 

c‘t DA},
- “35¢;
7 M u.\.6% ‘3 ‘
A». “H 2" .‘
Jr.

- t:
”131:" ~92ch 3'
§L1 ﬁ‘ygé‘“
. ﬁg ‘ ‘

. s

max. 15““ ,
.c . ..,:...;$” ”35%;" ' .- .‘ '3. ‘ a"
_ "‘ '31" .‘ ~ ‘ g ‘
'-‘ (W ”t“ “a “at v a?“ ‘ a 1 #35
.. . ﬁpﬁ 5..“ x- ! .. .4

A w v 7*
93:19::- ‘ ”- J

' ‘ 55a
3.2} ..

. “,3“?
.u VQ V‘.
ti-Efrr: ~

{:x '

.u.._

«£32552.
ﬁisﬁc’v ..2
' ‘ ”7:. 1‘: *2). w». .. ,9 .t.
..~@~.~.:—‘~.~:£x¢.-x *3 r -);q-.:,n: .
25H; ﬁ‘F'un‘. ,r-Ja'u . ‘ “at.
a. " ..--::.. .33" ‘7'
V 9“”, ' ' "
“p5" .

as.

2.1 Lake:
1‘51ng

~.
. sag“, (“ﬁr-vﬂ‘-’
L t“ a); I: " :1- x"?
> - {54$ “5‘ “3%? 5195‘
- v.9 *-

~5®L J-L

— I :- N‘ﬂ-I h

- :LVQ‘iy «L

‘l 3 x
; 1‘“ 132‘. :1;
. 3533‘“ . JA

., y“. o

. I w

$343141
- ”AV-x- NT
.3 " x 1 I

1..
v-\ (Ff

'I:r‘

 

: A
4 L; ' .'.‘a-ho

”'99. .

'révu‘xé'ﬁ".

1‘ .1”

Me'sasﬁﬂg' ' * my ..
' - . ‘ . ”v r .

New“ ., ., W ‘3,- ~%-' “L” .' .u;

‘ '35:..‘1'75k

“W"

“A“. v ‘ 7»:

W

i V “I “‘y~&"€$; "‘0 'ﬁ‘ . - - . -
«aw? .- MW : - . ~ -. - - .
amt? ﬂ» $2253: ’53th "é ‘ j - , ,- ragfgg

 

)HI TATE UNIVERSITY LIBRARIES

H Minimum/HIM!HINHHHIHII! III

will] LN" W

 

 

 

 

This is to certify that the

dissertation entitled
Representational Models of Attachment:

hplication and Reorganizational Processes
presented by

Lisa Elaine Crandell

has been accepted towards fulﬁllment
of the requirements for

 

 

Ph.D. degree in Clinical Psychology

Majorp e r

Q/«M £7 F
//
Date 7j6/‘Z/

MSUis an Afﬁrmative Action/Equal Opportunity Insmution

042771

 

LIBRARY
Michigan State
University

 

 

 

PLACE N RETURN BOXto mouthi- mum your need.
To AVOID FINES Mun on or before date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

I
D

 

 

 

II
!

 

 

 

I
E

 

 

 

 

 

 

 

 

 

Ell
III
III

 

 

 

MSU ltAn Afﬂnndlv. Action/Emil Opponunly lmwon

!

REPRESENT ATIONAL MODELS OF ATTACHMENT:
REPLICATION AND REORGANIZATIONAL PROCESSES

By

Lisa Elaine Crandell

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

1994

ABSTRACT

REPRESENT ATIONAL MODELS OF ATTACHMENT:
REPLICATION AND REORGANIZATIONAL PROCESSES

By

Lisa Elaine Crandell

This study examined the replication of dysfunctional parent-child
relationships across generations, the impact of maternal attachment security on
child outcome, and experiences underlying discontinuity in developmental
trajectories. Subjects were 46 mother-toddler dyads recruited from a community-
based sample. Utilizing an adaptation of the Adult Attachment Interview
(George, Kaplan & Main, 1985), the results indicated that the manner in which
childhood histories were mentally organized and integrated in adulthood imparted
a profound inﬂuence on parent-child behavior patterns in the next generation.
Specifically, mothers classified as secure expressed more warmth, affection and
enjoyment when playing with their children and they facilitated more self-reliant
behavior from their children than insecure mothers. Children of secure mothers
were more compliant and rated as more likeable than children of insecure
mothers. Additionally, the secure mothers and their children engaged in more
synchronous interactions than insecure dyads. Conversely, insecure mothers were
more hostile, negative and sarcastic in their affect. They were more intrusive,
manipulative and overcontrolling in their play and/or were more likely to engage
in emotionally disconnected, parallel play. The children of these mothers

expressed more anger, dislike or hostility toward their mothers, demonstrated

more emotional distress, and were more disobedient than children of secure
mothers. With respect to child outcome, children of secure mothers
demonstrated higher intellectual functioning and were more likely to be classified
as securely attached compared to children of insecure mothers. With regard to
the reorganization of attachment representations, virtually all mothers who had
experienced unloving and insecure parent-child relationships in childhood but
presented with secure attachment representations in adulthood (i.e., resolved) had
experienced an alternate attachment figure in childhood or adolescence whereas
almost none of the insecure mothers had experienced this type of relationship.
Additionally, the resolved mothers participated in individual psychotherapy in
adulthood for an average of two years compared to insecure mothers who were
not as likely to enter into individual psychotherapy and for those who did, the

average length of time of treatment was three to six months.

THIS DISSERTATION IS DEDICATED TO MY SISTER

MARSHA ANN CRANDELL BECK

THROUGHOUT MY LIFE

SHE HAS GIVEN NIB THE NU RTURANCE AND SUPPORT

THAT HAS ENABLED ME TO ACHIEVE THIS GOAL

iv

ACKNOWLEDGMENTS

I would like to express my appreciation and gratitude to my husband, Jim
Rigas, for the tremendous patience and support he has provided throughout my
graduate education. I wish to thank my companions, Ann Thompson and Laura
Symonds, for their support, encouragement, and endearing friendships. A
heartfelt thanks to my friends, Carol Misseldine and Ed Groves, who have taught
me much about the richness and complexity of relationships. I am also indebted
to Candy Kokinakis and Yogi Amrit Desai for their spiritual guidance.

I would like to acknowledge the contribution of my assistants Kathryn
Buysee, Anne Cupp, Lil Cybulski, Kandi Jones, Brad Klemm, and Bridget
Kramer, with a very special thank you to Jill Platte for her invaluable effort in
this project. I would also like to extend my gratitude to the families who
participated in this study. I am inspired by their courage and willingness to share
their lives.

Finally, I wish to express my deepest appreciation to my chairperson and
mentor, Dr. Hiram Fitzgerald. He has managed the very difficult task of
simultaneously supporting and challenging me. His guidance has sharpened my

skills and critical thinking and strengthened my commitment to research.

TABLE OF CONTENTS

List of Tables .............................................. viii
List of Figures ............................................... x
CHAPTER ONE

Literature Review ............................................. 1
Introduction ................................................. 1
Attachment Theory ............................................ 2
Attachment as a Biological Phenomenon ............................ 2
Internal Working Models ........................................ 4
Attachment Security and Interpersonal Relatedness .................... 9
Attachment Security and Parenting ............................... 10
Attachment Security, Peer and Intimate Relationships .................. 13
Modes of Interpersonal Relatedness ............................... 17
Summary of Theoretical Assumptions .............................. 20
Origins of Attachment Theory ................................... 20
Mother-Infant Interaction Studies ................................ 22
Intergenerational Replication of Attachment Security .................. 26
Summary of Parent-Child Interaction Research ....................... 29
Reorganizational Processes ..................................... 30
Recovery and Discontinuity of Attachment Security .................... 32
Goals and Hypotheses of Current Study ............................ 36
Rationale .................................................. 36
Hypotheses ................................................. 43
Hypothesis 1 ................................................ 43
Hypothesis 2 ................................................ 43
Hypothesis 3 ................................................ 44
Hypothesis 4 ................................................ 45
Hypothesis 5 ................................................ 45
CHAPTER TWO

Method ................................................... 48
Subject Selection ............................................. 48
Subject Description ........................................... 49
Measures and Procedure for Phase I .............................. 50
Measures and Procedure for Phase II .............................. 54

Table of contents (cont’d)

CHAPTER THREE
Results .................................................... 64
Hypothesis 1 ................................................ 64
Hypotheses 2 and 3 ........................................... 67
Supplemental Analyses ........................................ 84
Hypothesis 4 ................................................ 86
Supplemental Analyses ........................................ 94
Hypothesis 5 ................................................ 99
Summary of Results ......................................... 101
CHAPTER FOUR
Discussion ................................................ 108
Methodological Considerations ................................. 108
Replication Processes ........................................ 115
Reorganizational Processes and Residual Effects ..................... 121
Intelligence ................................................ 127
Limitations of the Study and Directions for

Future Research .......................................... 128
APPENDIX A ............................................. 131
APPENDIX B ............................................. 137
LIST OF REFERENCES ..................................... 142

vii

Table 1
Table 2

Table 3
Table 4
Table 5
Table 6
Table 7
Table 8
Table 9
Table 10
Table 11
Table 12
Table 13

Table 14

Table 15

Table 16

Table 17

LIST OF TABLES

Summary of Measures ............................... 63
Inter-rater Reliabilities for the Adult

Attachment Questionnaire ............................ 65
Intellectual Functioning by Adult Attachment

Questionnaire Classifications ........................... 66
Parenting Stress Index Comparisons of Mothers with

Resolved, Secure and Insecure Attachment ................ 68
Demographic Variables of Families with Resolved and

Secure Maternal Attachment ........................... 70
Parenting Stress Index Comparisons of Mothers with

Resolved and Secure Attachment ....................... 71
Parent-Child Interaction Patterns by Resolved and

Secure maternal Attachment ........................... 72
Parent-Child Interaction Patterns by Combined Secure

and Insecure Maternal Attachment ...................... 75
Intercorrelations Among Child Intellectual

Functioning and Attachment Story Completion Task .......... 78
Child Attachment Security by Resolved, Secure and

Insecure Maternal Attachment ......................... 80
Child Attachment Security by Combined Secure and

Insecure Maternal Attachment ......................... 82
Child Intellectual Functioning by Resolved,

Secure and Insecure Maternal Attachment ................. 83
Significant Differences in Demographic Variables for

Mothers with Combined Secure and Insecure Attachment ...... 85

Multiple Regression of Demographic Variables and

Maternal Attachment Security on Parent-Child

Interaction Patterns ................................. 87
Alternate Attachment Figure and Psychotherapy

Experience by Resolved and Insecure Maternal

Attachment ....................................... 88
Intercorrelations Among Child IQ, Mother IQ,

Alternate Attachment Figure, Psychotherapy,

and Maternal Attachment Security ...................... 91
Multiple Regression of Demographic Variables,

Maternal IQ and Maternal Attachment Security

on Child IQ ....................................... 95

viii

List of Tables (cont’d)

Table 18 Intercorrelations Among Parenting Behavior,

Child IQ and Parent IQ .............................. 97
Table 19 Parent-Child Interaction Patterns by Combined

Secure and Combined Insecure Maternal Attachment ........ 102
Table 20 Child Intellectual Functioning and Attachment

Security by Maternal Attachment Security ................ 104

LIST OF FIGURES

Figure 1 Predicted Model of Replication and Reorganizational
Processes of Attachment Representations .................. 47

Figure 2 Intercorrelations Among Alternate Attachment
Figure (AAF), Psychotherapy, Maternal IQ,
Maternal Attachment Security and Child IQ ............... 92

CHAPTER ONE
Literature Review

mm

In the past decade, there has been considerable interest and research
related to experiences in one’s family of origin and the impact of these
experiences on subsequent development. In particular, there have been numerous
studies investigating the replication of dysfunctional parent-child relationships
across generations (Crowell & Feldman, 1989; Fonagy, Steele & Steel, 1991; Haft
& Slade, 1989; Jacobvitz, Morgan, Kretchmar & Morgan, 1991; Pratt, Roth,
Cohn, Cowan & Cowan, 1991; Quinton & Rutter, 1988). Attachment theory
provides a rich framework for conceptualizing and exploring these issues of
continuity because it focuses on the emotional dimension of parent-child
relationships and the affective correlates of these early experiences in subsequent
relationships.

This study sought to advance understanding in the field of attachment
theory and parent-child relationships in the following ways: 1) by investigating
variations in current methodology for classifying representational models of
attachment; 2) examining the interplay among adult attachment security, parent-

child interactions, and child functioning in pre-school age children; and

2
3) exploring the role of alternate attachment figures, psychotherapy and

intelligence in the reorganization of attachment representation in adulthood.

W

There are four fundamental assumptions underlying classic attachment
theory: 1) that attachment is a biologically based phenomenon which serves to
facilitate the survival of the species; 2) that on the basis of the primary
attachment relationship, the child develops internal representations of self and
other; 3) that these representations organize the child’s affective experiences; and
4) that these self/other representations are the foundation for subsequent
relationships throughout development. The following review is organized around

these four assumptions.

Attachment as a Biological Phenomenon
Bowlby (1969/1982, 1973, 1980) proposed an ethological theory of

attachment which maintained that all infants have a biological propensity to
become attached to their primary caregiver(s). The function of this attachment
relationship is to protect the infant from environmental threats and dangers. To
this end, the child is biologically equipped with behaviors to elicit parental
attention and proximity in times of distress. Bowlby argued that diverse behaviors
such as crying, clinging, smiling, vocalizing, and following promote proximity with
the parent and thereby represent an "attachment behavioral system." Many of

these same behaviors have been observed in the animal kingdom where they serve

3
the same function. Thus, within this evolutionary framework, the attachment

relationship facilitates survival of the species by protecting the infant from harm.
There are also intra- and inter-personal sequelae emanating from the
attachment relationship. When the caregiver is available and responsive to the
child’s distress signals, the child learns the caregiver will reliably afford protection,
which provides the child with a sense of safety and security. It also validates the
child’s experiencing. In turn, this leads to the ability to differentiate self from
other. As this sense of security and differentiation evolve, there is increasing
physical and psychological separation between parent and child as the child
progressively explores and interacts with the larger world. Therefore, at a very
basic level, the message being communicated to the child, via the attachment
relationship, relates to the child’s connection with and separateness from others.
Franz & White (1985) summarize a number of theorists who argue that these
dimensions of connectedness and individuation are the fundamental aspects of
human experiencing. These two dimensions are inextricably intertwined in the
attachment relationship. As the parent responds to the child’s needs, the child is
provided with a sense of safety and security. This creates the experience of
connectedness that in turn provides a context for individuation. In this sense, the
early parent-child attachment relationship represents a foundational experience

for individual and interpersonal development.

Wo ' e

The second assumption of attachment theory is that the child develops
internal representations of self and other on the basis of the early attachment
relationship. Bowlby (1980) stated, "...a person who has formed a secure
attachment is likely to possess a representational model of attachment figure(s) as
being available, responsive, and helpful and a complementary model of himself as
at least a potentially lovable and valuable person" (p.242). Thus, a securely
attached child develops a fundamentally positive image of self and other. In
contrast, an infant who experiences the primary attachment figure as rejecting,
develops representations of the world as "comfortless and unpredictable" and of
him or herself as not worthy of help and comfort (Bowlby, 1980). Consequently,
an insecurely attached child develops a fundamentally negative image of self and
other. Bartholomew (1990) presents a model of self/other representations and
Bartholomew & Horowitz (1991) provide empirical support for the presence of
these self/other representations in two studies examining adult attachment styles.
Memories of secure attachment with one’s parents in childhood were associated
with a positive image of both self and other in adulthood. Memories of insecure
attachment were associated with three different patterns: a negative image of self
and other; a positive image of self and a negative image of other; or, a negative
image of self and a positive image of other. Although Bowlby had not predicted
these "mixed models", these authors argue that they are complementary models as
they relate to the individual’s self-worth and how he or she strives to either

protect or affirm this self-image in interactions with others. In addition, they

5

found that these combinations of self/other representations were significantly
related to issues of dependency, avoidance, and difﬁculties in intimacy in close
relationships in adulthood.

This lends support to the third assumption of attachment theory which
posits that self/other representations serve to organize the child’s affective
experiences throughout development. Although Bowlby did not articulate this
organizational process specifically, extrapolating from the theory generates the
following scenarios. On a global level, there are three basic interactional patterns
between parent and child: (1) the parent is sensitive to and fulfills the child’s
needs (nurturance); (2) the parent does not respond to or fulfill the child’s needs
(neglect); (3) the parent covertly or overtly dismisses and/or overrides the child’s
needs (rejection, and in extreme cases, abuse). When the caregiver responds
appropriately and consistently to the child’s cues, the child’s internal experience is
that "someone is there for me". This leads to a learned contingency, "When I am
uncomfortable, I can communicate my distress and it will be alleviated.” From a
classical conditioning paradigm, the infant also develops positive affect toward the
object relieving this discomfort so that felt security and emotional connectedness
converge with respect to the caregiver. The child develops a basic sense of trust
(Erikson, 1963), security, self-awareness, and confidence in his or her ability to
relate to and explore the world. This is secure attachment. There is a balance in
the child’s sense of connectedness to and differentiation from others.

When the caregiver is consistently rejecting or neglecting, the child

develops a sense of mistrust and insecurity because emotional needs are not

6
gratiﬁed. Additionally, experiencing unmet needs inspires anger and frustration.

Therefore, the child develops negative affect toward the caregiver. Main (1981)
argues that parental rejection and the child’s resulting anger are the central
determinants of avoidant attachment. She states that the function of the avoidant
pattern is to suppress emotion "as a conditional strategy for maintaining proximity
under conditions of maternal rejection" (p. 686). In other words, under distress,
the child’s attachment system is activated but previous attempts to seek comfort
and reassurance from the primary attachment ﬁgure have either been rejected or
ignored. The child is then faced with an irresolvable conﬂict. Avoidance
represents a moderately successful adaptation to this dilemma. The "shift in
attention" away from the attachment ﬁgure allows the child to maintain needed
proximity in threatening circumstances by minimizing rejection and curtailing
anger toward the caregiver. The child’s resulting anger and hostility from having
these attachment needs frustrated are then expressed in less threatening
situations. In fact, even Ainsworth, Blehar, Waters & Wall (1978) noted that
infants identiﬁed as avoidant in the Strange Situation often displayed aggression
and hostility toward their mothers during home observations. Within this context,
the child must learn to depend on him or herself, or devise ways of securing
protection from the caregiver. In both cases, this fundamental parental task
becomes the child’s own responsibility. This premature self-reliance may account
for the preoccupation with issues of independence, personal control, and an
inﬂated sense of self, which are consistently noted in avoidant adults.

Furthermore, these traits would restrict the capacity for emotional vulnerability

7

and intimacy in interpersonal relationships, which is another characteristic of
these individuals as they progress developmentally. Thus, avoidant attachment
represents an imbalance in human experiencing for the realm of individuation is
pursued at the expense of connectedness to others.

Ambivalent attachment would develop in response to inconsistent
parenting. While parental nurturance, neglect, and rejection are the global
patterns, there are variations and vacillations in these responses. Some parents
may be consistently rejecting while others may alternate between periods of
nurturance and rejection. In these cases, the child experiences sporadic
gratiﬁcation of emotional needs. Therefore, like the avoidant child, the
ambivalent child develops a general sense of mistrust and insecurity from
experiencing unmet needs and anger and frustration toward the caregiver for
failing to fulﬁll these needs. However, the caregiver is sometimes sensitive and
responsive to the child’s needs so like the secure infant, the ambivalent child also
develops positive affect toward the caregiver. The clinging, dependent behaviors
in conjunction with the angry, resistant behaviors accurately reﬂect the insecurity
and sporadic relief the ambivalent child experiences with the caregiver. Role-
reversal represents a speciﬁc type of parental inconsistency and is often found to
be related to ambivalent attachment. Parents who are role-reversing are
responding to their child on the basis of the parent’s own needs. Hence, rejection
of the child’s needs is implicit in role-reversing relationships. However, the
nature of the parent-child relationship in role-reversal is qualitatively different

from rejection. The parent necessarily pushes aside the child’s needs in the

8

process of imposing his or her own needs onto the child. Yet, because the child
is required to provide love and nurturance to the parent, their relationship is
often close and affectionate. The rejection characteristic of avoidant attachment
is a cool aloofness on the part of the parent and these parents often recoil from
physical affection with their children. Therefore, role-reversal represents a
speciﬁc type of parental rejection which is juxtaposed with experiences of physical
affection and closeness. In fact, physical and emotional proximity are maximized,
sometimes to the point where personal boundaries are blurred. In extreme cases,
the child is psychologically merged with the parent, which not only denies the
child’s needs, it precludes the development of his or her sense of self. Therefore,
in some combination, the underlying dynamic to ambivalent attachment is
parental inconsistency. When parents respond sporadically to their child’s needs,
the child develops a pervasive sense of mistrust and insecurity, a poorly developed
sense of self, and uncertainty about his or her ability to relate to the world. In
cases of role-reversal, the parent does not vacillate between periods of nurturance
and rejection, these conﬂicting experiences exist simultaneously and are woven
into the very fabric of the parent-child relationship. In both cases, ambivalent
individuals demonstrate signiﬁcant difﬁculties in the process of separation-
individuation and often remain entangled in the parent-child relationship, even
into adulthood. This vacillating struggle to individuate is the characteristic
feature of ambivalent attachment in adulthood. Fundamentally, the ambivalent

pattern represents the opposite adaptation as the avoidant pattern for these

9

individuals are invested in being connected to others at the expense of
individuation.

In summary, attachment security is conceptualized to develop in response
to parent-child interactional patterns along two dimensions: the quality and
consistency of the dyadic interchanges whereby quality distinguishes secure from
insecure attachment and consistency discriminates avoidant from ambivalent
attachment. From this context, the child develops representational models of self
and other which organize his or her affective experiences and orientation towards
others. This connection between attachment security and interpersonal

relatedness represents the fourth basic assumption of attachment theory.

Attachment Security and Interpersonal Relatedness

Given that the child develops internal self/other representations in
response to the attachment relationship, the fourth assumption of attachment
theory posits that the early parent-chﬂd relationship establishes the prototype for
subsequent relationships. Bowlby (1980) stated, "On the way in which an
individual’s attachment behavior becomes organized within his personality turns
the pattern of affectional bonds he makes during his life" (p.41). There is
compelling evidence that the mental model developed in response to early parent-
child relationships does inﬂuence the development of other signiﬁcant
relationships across the life span. The following sections will provide a review of

this research as it relates to parenting, peer, and intimate relationships.

10
c t ' a e ti

There are several studies which indicate that there is continuity between
adult attachment security and patterns of attachment developed with one’s own
children in adulthood. George, Kaplan & Main (1985) developed the Adult
Attachment Interview (AAI) which assesses the quality of an adult’s current
internal representations of parent-child attachment relationships from childhood.
The internal representations are evaluated by: the adult’s ability to access and
integrate early childhood attachment experiences with a general understanding of
the parent-child relationship; an awareness of the impact it has had upon their
personality and adult relationships; and resolution of anger toward parents. From
the AAI, three patterns of adult attachment were identified: secure, dismissing
and preoccupied. Individuals classiﬁed as securely attached had very diverse
experiences with respect to early parent-child relationships. Some individuals
experienced loving, secure relationships while others experienced physical and
emotional abuse, extreme role-reversal, psychological and physical neglect and/or
rejection. However, as adults, these individuals shared a common organization of
thought regarding these relationships: they had access to speciﬁc childhood
memories; valued early attachment experiences; were able to integrate positive
and negative aspects of their parents into a coherent view of the overall
relationship; and acknowledged the inﬂuence of these experiences on their
personality. Individuals classiﬁed as dismissing had difficulty accessing speciﬁc
childhood memories, devalued the importance of early attachment experiences,

and/or discounted the inﬂuence of these experiences on current functioning.

11
They were restricted in their affect when relaying events and were preoccupied

with issues of independence, personal strength, and control. Childhood
experiences typically involved either overt or covert parental rejection and an
absence of affection. In addition, their autobiographical memories often
contradicted their semantic or overall representations of parent-child
relationships. For example, some of these individuals responded that their
parents were loving and always there for them while relaying explicit incidents of
parental rejection. Those not invested in idealizing their parents may have
presented and discussed negative aspects and experiencing of their parents, but
quickly dismissed the impact upon themselves. Individuals classiﬁed as
preoccupied had access to speciﬁc memories, in fact were often ﬂooded by
negative experiences, but were unable to integrate these experiences into a
coherent understanding of the parent-child relationship. Their protocols were
marked by a sense of confusion and vacillation between the negative aspects of
their relationship with their parents and a continuing desire to please them.
Some of these individuals were preoccupied with anger and resentment toward
unloving parents and were engaged in an active struggle to individuate while
others had adopted a more passive stance. This passive subgroup had difﬁculty
separating parent from self and/or past from present, were vague and evasive in
their descriptions, and presented a weak sense of personal identity. Typically,
preoccupied individuals had experienced moderate to extreme role-reversing or
enmeshed parenting and all of them were still psychologically entangled in the

parent-child relationship.

12
From the standpoint of a mental model, the AAI identiﬁed three patterns

of organization of thought with respect to attachment relationships which are
consistent with the patterns of attachment identiﬁed by behavioral observations in
early childhood (Ainsworth, et a1, 1978). Using the AAI, several subsequent
studies demonstrated a link between adult attachment representations and child
attachment security. Main & Goldwyn (1984) note that mothers who were
labeled securely attached with respect to their adult representations of attachment
security were more likely to have children who were securely attached whereas
mothers who were classiﬁed as insecurely attached tended to have children who
were insecurely attached. Likewise, Main, Kaplan, & Cassidy (1985) found that
adult attachment security of mothers and fathers was signiﬁcantly related to the
attachment security of their infants. Similarly, several other researchers have
documented that insecure attachment representations from early parent-child
relationships are associated with insecure attachment relationships in second
generation parent-child dyads (Crowell & Feldman, 1989; Haft & Slade, 1989;
Jacobvitz et a1., 1991; Pratt et a1., 1991).

Collectively, these studies provide considerable support for the notion that
working models of attachment do inﬂuence subsequent parent-child relationships.
However, it is important to keep in mind that the AAI depends on retrospective
accounts of historical events and therefore, it is not clear to what extent the
reported memories of these events correspond to actual reality. As such, it is the
individual’s "current state of mind" with respect to attachment relationships that is

linked with the continuity of attachment security in the following generation.

13
These studies provide partial support for Bowlby’s claim that parent-child

attachment relationships represent the prototype for affective bonds throughout
life. However, to fully support this position, the mental model constructed should
not be circumscribed to parent-child relationships, but should involve a more
global representation that encompasses other close relationships. The data on

peer and intimate relationships address this issue.

mchment Security, Peer, and Intimate Relationships

There are several studies which indicate that there is continuity between
attachment security and the quality of interpersonal relationships. Sroufe (1989)
reports that pre-school children elicited responses from teachers consistent with
their attachment histories. In general, teachers responded to the three groups as
follows: secure children received warm, positive gestures and high expectations
for compliance; ambivalent children were treated as more immature and needy
with low expectations for compliance; and avoidant children were less likely to
seek teacher support and were the only children to elicit teacher anger. Similarly,
Pastor (1981) observed dyadic interactions in preschool children who had
previously been classiﬁed by the Strange Situation. He found that the interactions
of the dyads involving secure/avoidant and ambivalent/avoidant children were
more aggressive and hostile than the secure/secure, secure/ambivalent, or
ambivalent/ambivalent dyads. In addition, the interpersonal hostility was
reciprocated by both members in the avoidant dyads. These studies indicate that

children were involved in interactions with teachers and peers that were consistent

14
with their attachment histories, underscoring the power of parent-child

relationships in shaping the nature of other interpersonal relationships in
childhood.

There are also several studies suggesting continuity between attachment
representations of parent—child relationships from childhood and the quality of
intimate relationships in adulthood. Kobak & Sceery (1988) interviewed ﬁrst-year
college students using the AAI, obtained Q-sort ratings from peers regarding
these individuals’ social interactions, and assessed self-perceptions on the basis of
questionnaires. They discovered signiﬁcant differences in interpersonal styles that
were related to adult attachment security. Speciﬁcally, they found secure
adolescents were rated more positively by peers and reported feeling more
content and supported in their social relationships than insecure adolescents.
Preoccupied individuals were rated as very anxious by peers and reported feeling
high levels of personal distress but experienced some support in their social
relationships. Dismissing individuals were rated as hostile by peers and reported
feeling very distant in their social relationships.

Hazan & Shaver (1987) examined the relation between adult attachment
security and romantic love styles. On the basis of questionnaire data, they
identiﬁed secure, avoidant, and ambivalent attachment patterns and accessed
information regarding adult romantic relationships. Individuals classified as
securely attached described warmer relationships with their parents and between
their parents in their early childhood experiences than insecurely attached adults.

They reported their adult love relationships were based on trust and friendship,

15

experienced them as more satisfying, and were engaged in more enduring
relationships than avoidant or ambivalent adults. Individuals classiﬁed as
avoidant generally described their parents, particularly their mothers as being cold
and rejecting. They indicated that their adult relationships were characterized by
lack of trust, fear of intimacy, and moderate vacillation in emotional highs and
lows. Individuals classiﬁed as ambivalent described their parents as being
intrusive, unfair, and threatening and the parental relationship was characterized
as unhappy. They indicated that their adult relationships were marked by
obsessive preoccupation and sexual attraction, desire for union and intense
jealousy, and extreme emotional highs and lows.

Cohn, Silver, Cowan, Cowan & Pearson (1991) report that adult
attachment security, as assessed with the AAI, was not related to marital
satisfaction by self-report measures but was related to couples’ behavior in
romantic relationships. Speciﬁcally, couples comprised of two insecurely attached
partners engaged in more conﬂict and less positive interactions than couples with
two securely attached partners. Couples comprised of a secure husband and an
insecure wife resembled the couples with two secure partners. The authors
suggest that perhaps the presence of a secure partner has an ameliorative effect
on the quality of the relationship. Since there were only two couples composed of
a secure wife and an insecure husband, it is not known whether either partner can
ﬁll this role or whether a secure husband plays more of a stabilizing role in the

marital relationship.

16

Using questionnaire measures to classify adult attachment security, Feeney
& Noller (1990) report that secure adults were more trusting in their relationships
and were involved in longer and more stable relationships than insecure adults.
Ambivalent subjects expressed dependence and desire for commitment in
relationships yet their relationships were the least enduring. Avoidant subjects
reported feeling distant and mistrustful in intimate relationships and were more
likely to report never having been in love.

Finally, Collins & Read (1990) report adults were generally involved in
romantic relationships that conﬁrmed their working models. For example, most
subjects had romantic partners who shared similar beliefs and feelings about
closeness, intimacy and the dependability of others. However, some individuals
who were anxious about being abandoned and unloved chose partners who were
uncomfortable with closeness and intimacy. These authors echo Weiss (1982) in
suggesting that many individuals seek partners consistent with how their
attachment systems have been prepared to respond. Those individuals who expect
partners to be psychologically available and dependable, choose partners that
meet those expectations and vice-versa.

In terms of relationship satisfaction, Kobak & Hazan (1991) report that
adult attachment security was related to partner ratings of marital adjustment and
couples’ ability to maintain constructive communication during problem solving.
Finally, Frommer & O’Shea (1973) found that separation from family of origin in
childhood was signiﬁcantly related to marital conﬂict and parenting difﬁculties.

Although not directly assessing the continuity between adult attachment security

17

and romantic relationships, separation is a primary determinant in the
development of attachment and it is interesting to note that mothers who did
experience this disruption from their families of origin went on to have difﬁculties
in their emotional connections to their husbands and children.

These patterns of attachment identiﬁed in infants and toddlers by the
Strange Situation and in adolescents and adults by interviews and questionnaires
strongly resemble the interpersonal trends detailed by Homey (1950). A
summary of her theory and its relation to attachment theory is provided in the

following section.

Modes of Interpersonal Relatedness

Homey proposed that when caregivers are insensitive to a child’s
developing individuality and emotional needs, the child feels unloved and isolated.
This experience creates a sense of core insecurity or "basic anxiety". Attempts to
alleviate this basic anxiety are evidenced in three psychological solutions: the
expansive solution (appeal to mastery); the self-effacing solution (the appeal to
love and morbid dependency); and the solution of resignation (the appeal to
freedom). Individuals who adopt an "appeal to mastery" attempt to conquer their
fears and conﬂicts. They seek perfection and complete self-sufﬁciency, are
disdainful of others, and require admiration and blind obedience in an effort to
alleviate their own sense of inadequacy and inferiority. They are openly arrogant,
ambitious, aggressive and demanding. Often they are driven by vindictiveness,

revenge and a need to triumph over others. Their sense of being unlovable is

18

translated into anger toward others and bitter resentment. Individuals who solve
their internal conﬂict by adopting an ”appeal to love" dramatize their neediness
and helplessness. They alleviate their basic anxiety by appeasing and becoming
dependent on others. For these individuals, "love" is their cure. As Homey
writes,

"Others should supply the initiative, do his work, take the

responsibility, give meaning to his life, or take over his life so that

he can live through them...the power which the appeal of love has

for the self-effacing type becomes perfectly clear. It is not only a

means to allay anxiety; without love he and his life are without

value and without meaning. Love therefore is an intrinsic part of

the self-effacing solution...love becomes as indispensable for him as

oxygen is for breathing" (p.228).
Individuals who ”appeal to freedom" withdraw as a means of coping with their
inner conﬂicts. Horney describes this process of resignation as "a shrinking,
restricting and curtailing of life and growth" (p. 260). While the other modes are
characterized by an active reaching out or involvement with others, this mode is
characterized by detachment and an emotional distance from others. These
individuals engage in a life void of "pain or friction or zest". They purposely
avoid becoming too involved or attached to anyone or anything. Their
relationships and attachments are restricted to a level of experiencing that
prohibits a sense of neediness or dependency.

There is a thematic continuity between these interpersonal orientations
that Homey described as "moves toward, against and away from others" and

attachment styles. The "moving away" solution encompasses characteristics of the

avoidant/dismissing types whereas the "moving toward" profile characterizes the

19

ambivalent/preoccupied types. The "moving against" solution shares the anger
and hostility present in the avoidant type. However, the "moving against"
orientation also shares the vacillation between connectedness and disengagement
from others that is characteristic of the ambivalent type. In that sense, the
"moving against" orientation appears to represent a mix between the avoidant and
ambivalent attachment patterns. Horney stated that these modes are solutions for
intrapsychic conﬂict as well as determinants for the form of human relationships.
She suggested that the connection between these processes involves the
experiencing of love. The common dynamic is a deeply ingrained sense of feeling
unloved and hence, unlovable. Those who "move agains " respond with anger and
hostility at the injustice of feeling unloved as a child. Those who "move toward"
desperately seek to attain affirmation of their lovability while those who "move
away" protect themselves by dissociating from experiences of love and attachment.
The common outcome is rigidly relating to others via these modes and distorting
the experiencing of interpersonal relationships to accommodate these orientations.

While there is no direct evidence for Horney’s model, the theoretical
overlap between the modes of relating she described and the styles of attachment
derived from attachment theory are remarkably similar. In addition, both
theoretical models postulate that experiences of parental love and nurturance
become incorporated into beliefs about the self and are subsequently translated

into characteristic ways of relating to others.

20

W

In summary, the underlying assumptions of attachment theory are as
follows: children are biologically predisposed to develop an attachment
relationship with their primary caregiver(s); they develop internal representations
of self and other based on this relationship; these representations consolidate into
interpersonal styles of relating that can be observed as early as 12-24 months of
age; and that these styles appear to be somewhat stable throughout development.

The theoretical and empirical efforts presented thus far provide support
for these assumptions. They all converge on the idea that the emotional
relationship between parent and child has a signiﬁcant impact upon subsequent
individual development and interpersonal relatedness. However, the studies
represent diverse samples, measures, and content areas, making it very difﬁcult to
know to what degree these results can be generalized. In addition, it still remains
unclear what inﬂuences the child brings to the parent-child relationship, what
specific parent-child interactions underlie and differentiate secure and insecure
infant attachment, and exactly how these patterns are related to the reiteration of
attachment relationships throughout development. These issues are addressed in

the remaining sections of the literature review.

Origins of Attachment Security

While there is still lively debate about the role of infant temperament in
the development of attachment, there is growing consensus that infant

temperament may contribute to the expression of attachment behaviors but is not

21

a causal factor in the development of secure or insecure attachment (See Bclsky
& Isabella, 1988 for an in-depth review of this literature). For example, an
irritable infant may cry more often and be more difﬁcult to soothe but as long as
the parent is sensitive and responsive to these qualities, the child will have the
experience of being cared for that evolves into a secure attachment. In this way,
the caregiver is assumed to be the dominant inﬂuence in determining the quality
of the attachment relationship. There is some indirect support for this position in
two studies that demonstrated the predictive power of adult attachment security.
Benoit, Vidovic, & Roman (1991) and Fonagy et a1., (1991) used the AAI and
interviewed mothers during pregnancy. A 77% and 75% concordance rate was
reported between maternal attachment security during pregnancy and child
attachment security at 12 months of age, respectively. Although child
temperament was not assessed, the data indicate that for the majority of women
in these two studies, child attachment security was determined by maternal
functioning rather than characteristics of the infants.

It remains a possibility, however, that there is an interaction between infant
temperament and parental attachment security. Parents with secure attachment
representations may be capable of providing sensitive and contingent caregiving to
a wide range of infant temperaments. Parents with insecure attachment
representations may be able to provide this type of caregiving to infants with
"easy" temperament styles but the challenges of more "difficult" temperaments
may trigger the unresolved issues from their own histories in a way that prevents

them from responding to their infants. While the potential for this interaction

22

effect remains open for debate, there are numerous studies that have examined

the role of parental behavior in the development of infant attachment security.

Mother-Infant Interactions Studies

On a global level, many studies have identiﬁed parental ”sensitivity" and
"responsiveness" as primary determinants in the development of attachment
security (Ainsworth et a1., 1978; Belsky, Rovine & Taylor, 1984; Belsky &
Isabella, 1988; Benn, 1986; Egeland & Farber, 1984; Lamb & Easterbrooks,
1981). But there have only been a few studies to investigate ﬁner bands of
behavior within these global constructs. Studies that have addressed this area
emphasize the affective quality of parent-child interactions as a root variable in
attachment security.

Studying mother-infant interactions, Goldberg, MacKay-Soroka &
Rochester (1993) found that maternal responsiveness to infant affect was
associated with secure, avoidant and ambivalent infant attachment. Secure infants
demonstrated equal displays of negative, positive and neutral affective events and
mothers of these infants were responsive to all three types of affective displays.
Further, these mothers provided the most comments and verbal elaborations
regarding emotions. Conversely, avoidant infants displayed the least amount of
all three affective events, and their mothers were the least comforting and
responsive to displays of affect, particularly to negative emotional expressions.
Also, mothers of avoidant infants provided the fewest comments and verbal

elaborations about emotions in general. Ambivalent infants demonstrated

23

primarily negative affect and their mothers were most responsive to negative
affective events. Like the mothers of avoidant infants, these mothers provided
few comments and dialogue about the presence of emotions in general.

Izard, Haynes, Chisholm & Baak (1991) note that in a longitudinal study of
mother-infant dyads, maternal emotional experiences and expressiveness during
infancy predicted infant attachment security at 13 months of age. Speciﬁcally,
mothers of secure infants were more emotionally expressive with regard to
positive and negative affect. Mothers of insecure infants reported experiencing
more negative emotions and less positive emotions than mothers of secure
children, yet they reportedly suppressed their expressions of negative emotions
around their children. Ironically, their children demonstrated more anger,
sadness and distress in mildly stressful conditions than secure children.

In studies with mother-infant dyads at three and six months of age,
Tronick, Ricks & Cohn (1982) found complex relationships between infant
behavioral responses and maternal expressions of affect that were subsequently
related to infant attachment security at 12 months of age. In conditions of
"normal" face-to-face parent—infant interactions, mothers were rated along three
dimensions: 1) the extent to which they were responsive to the infant’s behaviors
by imitating or exaggerating infant social actions ("elaboration"); 2) the extent to
which they intruded on the infant’s activities ("overcontrolling"); and 3) the extent
to which they were hesitant and/or withdrew during the interaction
(”undercontrolling"). Following this condition, mothers were instructed to

"distort" mother-infant interactions by simulating depressed affect. Mothers who

24

scored higher on ”elaboration” during normal interactions had infants who
attempted to use positive social behaviors to reengage their mothers in the
depressed condition. Mothers who scored higher on "overcontrolling" and
”undercontrolling" had infants who either used negative social behavior or
withdrew from the interactions with no attempt to reengage their mothers in
interaction. How the infants adapted to the depressed condition was predictive of
their attachment security at 12 months of age. An overwhelming majority of the
infants who used the positive strategy were classified as securely attached at 12
months of age, whereas infants who used the negative or withdrawal strategies
were classiﬁed as insecurely attached. Of further interest, there was a dramatic
increase in infant distress and disjointed cycles of positive, negative and neutral
affective states for all infants during the depressed condition compared to the
normal interaction sequences. These authors present observable data that
demonstrate: 1) infants are remarkably sensitive to maternal behavior and affect;
2) maternal style of interaction and expression of affect is directly related to
differential organization of infant behavior; and 3) these strategies become
stabilized over time and are used for negotiating subsequent stress.

Isabella & Belsky (1991) provide further evidence for this notion of
"disjointed cycles" in parent-infant interactions as it relates to attachment security.
They report in a longitudinal study of mother-infant dyads at 3 and 9 months of
age that reciprocal and mutually rewarding behavioral interactions were predictive
of infant attachment security at 12 months of age. Speciﬁcally, mothers who were

sensitive to infant behavioral and affective cues by providing appropriate and well-

25
timed responses had infants who were later classiﬁed as securely attached.

Conversely, mothers who were intrusive and overcontrolling had infants classiﬁed
as avoidant, whereas mothers who were unresponsive and/or inconsistent had
infants classiﬁed as ambivalent at 12 months of age. These authors conclude that
the experience of dyadic synchrony that evolves from appropriate and predictable
parent-infant interactions is the substrate of secure attachment by providing the
infant with "...an optimal degree of interactive stimulation as well as a predictable
interactive partner" (p. 381).

Additional studies highlight the role of parental rejection and lack of
emotional warmth in the avoidant attachment pattern. Based on in-home
observations, Ainsworth et a1., (1978) noted that mothers of avoidant infants
displayed an aversion to close bodily contact and rejected their infants’ attempts
for physical affection. Egeland & Farber (1984) report similar patterns of
parenting during in-home observations of mother-infant interactions during
feeding and play activity. Mothers of avoidant infants avoided physical contact
except when necessary, and even then responded in a stiff and mechanical
fashion. They were less responsive and effective in their response to infant crying
and were psychologically unavailable to their infant. Mothers of ambivalent
infants demonstrated poor caregiving skills, although they were more adept than
mothers of avoidant infants. Similarly, Tracy & Ainsworth (1981) report that
mothers of avoidant infants were less likely to bug or cuddle their infants than
mothers of either ambivalent or secure infants. In times of distress when the

attachment behavioral system is intensely activated, children seek close bodily

26

contact for reassurance. For avoidant infants, these efforts are rebuffed.
Collectively, these studies of parent-child interactions highlight the role of
emotional warmth, affect expression, and regulation in the mother-infant dyad
and suggest that these are core elements of attachment security that underlie

parental "sensitivity" and "responsiveness."

hgrgenegtional Replication of Attachment Security

The remaining questions to be addressed are what differentiates parents
within the affective realm and speciﬁcally how is it related to repetition processes
across generations? Since the Main & Goldwyn (1984) study, adult attachment
security has been consistently linked with infant attachment security (Benoit et a1.,
1991; Fonagy et a1., 1991; Jacobvitz et a1., 1991; Pratt et a1., 1991). These studies
point to the organization of adult attachment representations as the underlying
process linking attachment patterns across generations. The implicit theoretical
model of this transmission hypothesis is as follows. Attachment representations
provide a template for parent-child relationships and thus serve to structure
dyadic interchanges on a behavioral and affective level. Secure attachment
representations in adulthood reﬂect a ﬂexible, non-defensive organization of
thought that permit open, sensitive and contingent caregiving to a broad range of
child verbal and non-verbal behaviors. Insecure representations reﬂect distorted,
defensive strategies that create restrictive, intrusive and/or inconsistent parental
responses to child behaviors. Therefore, parents with secure attachment

representations are more aware of their own internal experiencing and have

27

access to a broader range of affective experiences which, in turn, enables them to
provide more sensitive and responsive caregiving to their children. Parents with
insecure attachment representations are restricted and limited in these domains
by virtue of their representational models. However, there have only been a few
attempts to investigate this transmission hypothesis using a behavioral level of
analysis.

In a cross-sectional study with mothers and their three-year-old children,
Crowell & Feldman (1989) found that maternal attachment security, as assessed
with the AAI, was associated with maternal and child behavior in semi-structured
play sessions. Speciﬁcally, they report that secure mothers had warm and
responsive interactions with their children and a style of assistance which
promoted learning and self-discovery. They provided clear instructions in tasks
requiring compliance and expressed enthusiasm in play activities. Preoccupied
mothers were confusing and inconsistent, and had difﬁculty getting compliance
from their children. In play activities, they vacillated between periods of warmth
and gentleness and periods of anger and frustration. Dismissing mothers were
directive and controlling, unsupportive, and lacking in closeness and physical
affection with their children. Additionally, children of the preoccupied and
dismissing mothers were more angry, anxious, disobedient, and/or subdued than
children of secure mothers.

Haft & Slade (1989) present data from a pilot study with mothers and one-
year-old infants that focused on the emotional interchanges between parent and

child. They found that mothers with secure attachment representations were

28

consistently responsive to their infants’ expressions of positive and negative affect.
Preoccupied mothers were randomly responsive to infant affective expressions,
frequently vacillating between ignoring and misinterpreting their infants’ cues.
Dismissing mothers consistently distorted and misinterpreted their infants’ cues in
a selective manner. Speciﬁcally, they rejected their infant’s expression of anger
and distress as well as their cues for comfort and reassurance. This rejection
initially involved an attempt to "override" the infant’s experience through a
counter comment and often progressed into ridicule and sadistic comments if the
infant persisted. These authors suggest that "moment-to-moment interchanges"
create a context of "affective sharing", and that through this experience, infants
incorporate maternal attachment representations as part of their own psychic
structure. When parents are "attuned" to their infants’ affective cues and
respond in a consistent manner, this structure reﬂects reciprocal and mutual
experiencing. This echoes the conclusion drawn by Tronick et al. (1982) that the
infant’s capacity for affect regulation and stabilization is derived from parent-child
interaction patterns. When this interactive process produces a, "...shared, positive
emotional state, the infant develops a sense of effectance out of their cumulative
repetition, but when such interactions do not accomplish this goal the infant
deveIOps a sense of ineffectance or helplessness" (p. 84). They posit that this
quality of competence or ineffectance is then carried forward into subsequent
interactions and relationships.

Jointly, the Crowell & Feldman (1989) and Haft & Slade (1989) studies

indicate that maternal attachment representations differentiate parenting behavior

29
along the dimensions of style of assistance and affective quality in parent-child

interactions. This is a signiﬁcant advancement in attachment research. While
previous research has demonstrated a link between maternal behavior and child
attachment security, these are the only studies to examine mother behavior on

this level of analysis as a function of maternal attachment security.

Summary of Parent-Child Interaction Research

The data presented from parent-child interaction studies provide
compelling evidence that the affective quality and patterning of parent-child
interactions around the negotiation and fulﬁllment of the child’s basic needs are
the substrate of the child’s mental organization of attachment. Furthermore, from
these experiences, the core psychological structures of self and other evolve and
become consolidated into modes of interpersonal relatedness that are carried
forward throughout development.

What has yet to be clariﬁed is precisely when these self/other
representations become consolidated, the impact of simultaneously competing
experiences on these representations, and the circumstances which permit and/or
facilitate their reorganization. This question regarding the potency of early
parent-child relationships is a core issue in attachment theory. The notion of
"internal working models" implies that an individual continually revises and
elaborates mental representations throughout the developmental process.
However, the theory also predicts that these models become increasingly stable

and resistant to change because incoming information is either restricted or

30
diston as it is assimilated into the existing organization. Thus, while
fundamental change of self/other models may be possible, altering these
representations should become increasingly difﬁcult as the child progresses
developmentally. Therefore, even though early parent-child attachment
relationships are assumed to represent the prototype for future relationships, it is
not theoretically or empirically clear at what point in the developmental process
these internal models of attachment security become consolidated or what
processes and variables facilitate their reorganization. These questions were a

central focus of the current study.

Reorganizational Processes

The studies reported earlier that examined continuity between adult
attachment security and patterns of child attachment in the following generation
provide substantial evidence that painful and unloving relationships with one’s
parents in childhood are strongly associated with the replication of these
experiences when parenting one’s own children. However, there were also
instances of discontinuity. While these instances were not studied systematically,
there were commonalities across studies. Resolving the difﬁculties stemming
ﬁ'om parent-chﬂd relationships in childhood appeared to be a critical variable for
healthy parenting in adulthood. Parents who had conﬂicted and/or unloving
relationships with their own parents in childhood but had come to terms with
these experiences in adulthood were able to break the pattern and foster secure

attachments on the part of their children whereas those parents who had not

31

resolved these issues perpetuated the pattern with their own children. This is
consistent with observations made by Fraiberg, Adelson & Shapiro (1975) who
note that mothers who dissociated themselves from the affect arising from their
own mistreatment were more vulnerable to repeating the pattern of parenting
with their children whereas connecting with childhood pain and anger freed
mothers from recreating the negative parenting cycle.

Ricks (1985) describes a series of studies investigating maternal attachment
history, current marital relationships, and parenting difﬁculties. Results of these
studies indicate that maternal rejection in childhood, in conjunction with
defensiveness and idealization of parents in adulthood, were significantly related
to insecure infant attachment. Maternal histories of disruption or rejection that
were associated with positive child outcomes occurred in the context of stable
marriages, positive self-esteem and strong bonds to in-law families. Ricks
suggests that what allowed these mothers to modify a maladaptive parenting style
was engaging in a "signiﬁcant emotional experience" which altered their self/other
representational models.

Collectively, these studies suggest that accessing childhood pain is a
powerful deterrent against repetition in parenting, while repression and isolation
of painful affect is linked with reiteration of these patterns with one’s own
children. Individuals who engage in this process and/or encounter a "corrective
emotional experience", are able to break the cycle of parenting experienced and
parenting provided. However, there are only a few studies that have directly

examined discontinuities from one’s family of origin to adult functioning.

32

e ad ' n'u‘ of ttachmet ecuri

QUinton & Rutter (1988) conducted studies with women who had been
placed in foster care since early childhood ("ex-care" group) and compared the
quality of their parenting and marital relationships in adulthood with women who
had been raised with their biological families. They found that overall, the "ex-
care" woman had poorer social functioning, more difﬁculties in parenting, and
more disrupted and conﬂicted marital relationships in adulthood. However, a
signiﬁcant number of these women were not having psychosocial difﬁculties and
demonstrated good parenting skills. What differentiated these women from the
other "ex-care" women was the presence of a non-deviant and supportive spouse.
In other words, the absence of a supportive spouse was not associated with
problems in parenting for the control group but was for the "ex-care" women.
The "ex-care" women were more vulnerable to parenting difﬁculties based on
their childhood experiences but the presence of a supportive spouse mediated this
link. Therefore, Quinton & Rutter suggest that choices and experiences in adult
life can have a tremendous impact on reducing or minimizing adversity from
childhood. Similarly, Wambolt (1991), found that for young adults with
problematic family of origin experiences, choosing partners from dissimilar family
of origin experiences provided a means of escaping from repetition of relationship
patterns. These two studies are consistent with Rick’s (1985) argument that a
"signiﬁcant emotional experience" in adulthood is a central element in redirecting

the course of one’s life. While this may be true, it is not clear why some

33
individuals from pathological family backgrounds are able to choose healthy and

supportive spouses when the majority of these individuals do not.

Crandell (1992) reports data from a pilot study which suggest that the
reorganization of attachment representations may be related to a "corrective
emotional experience" in adulthood, but the ability to engage in and assimilate
this experience is rooted in a secure attachment relationship with an alternate
attachment ﬁgure at some point in the person’s early developmental history. All
of the subjects in this pilot study had experienced severely unloving and/or
abusive relationships with both parents during childhood. Subjects were classiﬁed
as either secure or insecure based on their representational models of parent-
child relationships. Utilizing an interview which focused on attachment
relationships across the life span, she found that experiencing a secure attachment
relationship with some other adult during childhood differentiated women who
had reorganized their representational models of attachment in adulthood from
those who had not. Similarly, Kaufman & Zigler (1987) report that individuals
who had been physically abused by their parents during childhood and were able
to break this cycle with their own children had experienced a supportive
relationship with one parent during childhood. Quinton & Rutter also mention
that while the pathway associated with healthier outcome involved a complex
network of variables, many of the women who were able to redirect their lives

reported having experienced a loving caregiving relationship with someone during

chﬂdhood.

34
The qualitative data from the Crandell (1992) study suggest that the

experience of an alternate attachment ﬁgure can have a major impact on
self/other representations. This was evident in the following comments. When
asked what this relationship taught them, typical responses included:

"...to care about myself and let me think that I’m important -- that

I’m a worthy person...that I’m lovable"; "That someone can be there

for me, for me, as opposed to my mother or grandmother who

really wanted me more for them. And that was important...Probably

because I had that relationship, I had more of a sense of self"; and

"I used to think I was unlovable or like defective or something but

my sister loved me and took care of me...and so I thought maybe I

am lovable." ‘

Finally, although there is some data addressing the continuity of
attachment security classifications across time, it is largely restricted to early
childhood and is inconclusive. Several studies provide support for the cross-time
stability of attachment security in young children (Main, et a1., 1985; Main &
Weston, 1981; Sroufe, Fox & Pancake, 1983; Waters, 1978). However,
Thompson, Lamb & Estes, (1982) and Vaughn, Egeland & Sroufe, (1979) report
a 60% and 53% stability rate, respectively, in attachment classiﬁcation for
children in high-risk samples marked by poverty, divorce and frequent changes in
residence. These two latter studies suggest attachment security may change in
childhood in response to changes in the family and caregiving environments.
Bartholomew (1990) writes,

”In early childhood, models of the self and other are not expected to

be well formed and certainly not clearly differentiated. Rather it is

continuity in the quality of the family environment that appears to

be largely responsible for continuity in attachment-related behaviors.
It is only over the course of time that the quality of attachment

35

relations is integrated into stable and self-sustaining internal
representations" (p. 165).

However, there is no empirical support for the latter part of this statement since
there is little data on the cross-time stability of attachment classiﬁcation in
adulthood. Therefore, it is not known whether there actually is more plasticity in
attachment security during childhood compared to adulthood, or at what point in
the developmental process it may become more resistant to change. This issue is
crucial if we are to understand continuities and discontinuities in attachment
relationships across the life span. In the Crandell (1992) study, all of the women
who had reorganized their representational models of attachment had experienced
a loving caregiving relationship by adolescence whereas none of the insecure
women had. This suggests that attachment representations can be signiﬁcantly
inﬂuenced by relationships outside of the parent-child relationship but these
representations may begin to be consolidated by adolescence. Consequently, in
the absence of a loving caregiving experience by adolescence, it may be
improbable that an individual could utilize a "corrective emotional experience" in
adulthood that would shift their developmental trajectory. Additionally, it may be
increasingly unlikely that an individual would even experience such an encounter.
Several researchers have commented upon the tendency for individuals to select
and create later social environments that conﬁrm their expectations
(Bartholomew, 1990; Bartholomew & Horowitz, 1991; Collins & Read, 1990;
Sroufe & Fleeson, 1986). However, the data from the Crandell study were based

on a small number of subjects and are only preliminary. There is still much we

36

do not understand about the processes underlying the formation, consolidation,
and reorganization of attachment representations. The present study was

designed to explore these issues.

Goals and Hypotheses of Current Study
Rationale

There were three overarching goals governing this study: 1) to investigate
variations in current methodology for classifying representational models of
attachment; 2) to examine specific processes underlying the replication of
problematic parent-child relationships across generations; and 3) to explore
processes and experiences underlying discontinuity in developmental trajectories.

With regards to the ﬁrst goal, virtually all studies demonstrating continuity
in the quality of attachment security in parent-child relationships across
generations have utilized the AAI. While the validity of this instrument has been
well documented, it requires extensive resources to administer, transcribe, and
score. This has posed signiﬁcant limitations to research in this area. Therefore,
the present study examined whether the AAI could be converted into a
questionnaire (AAQ) and used for classifying subjects broadly as either secure or
insecure in their attachment representations. The AAQ contained basically the
same questions posed in the AAI. The major changes involved administering the
instrument as a questionnaire measure, as opposed to a one-on-one interview, and
using a modiﬁed version (Adams, 1992) of the Adult Attachment Rating and

Classiﬁcation System (Main & Goldwyn, 1989) for classifying attachment security.

37
Adams (1992) found that maternal responses on a truncated version of the AAQ

were meaningfully related to parental perception of child behavior. However, he
utilized a group administration in his study and he did not use the AAQ as a
means for classifying adult attachment representations. Additionally, while other
researchers have developed and used questionnaire measures to classify adult
attachment security (Collins & Read, 1990; Hazan & Shaver, 1987), these have
been check-off lists that require subjects to classify themselves. The validity of
these methods has been questioned given the tendency for insecurely attached
adults, particularly those of the avoidant type, to distort and misrepresent their
erqreriences. The present study was the ﬁrst attempt to combine a comprehensive
and in-depth instrument, comparable to a clinical interview, with the practical
administration of a questionnaire to classify adult attachment security. In
addition, the study sought to examine how representational models of parent-child
relationships from the past, as assessed by the AAQ, related to self-perceptions of
distress in second generation parent-child relationships.

With respect to the second goal, previous studies on parent-infant
interactions have provided an entrance for examining the interplay among adult
attachment representations, parent-child interactions, and the reiteration of
parent-child relationships patterns across generations. Yet, there is need for
empirical work to further elucidate the factors underlying these processes and
how they present at different developmental stages. For example, previous
studies have demonstrated that parental style of interaction (e.g., consistency and

degree of control), as well as the quality of parental affect that is experienced and

38

shared, are both powerful determinants in the child’s emotional development.
However, only the Haft & Slade (1989) study describes parental behavior along
both dimensions as a function of parental attachment representations.
Additionally, with the exception of Crowell & Feldman (1989), studies of such
dyadic interactions at a behavioral level of analysis have focused almost
exclusively on mother-infant pairs so that it remains unclear how these
interactional patterns and child attachment security present at later points in
development. A major obstacle to this research has been how to assess
attachment security in older children since proximity seeking behaviors are no
longer activated during physical separations from parents. As children advance in
cognitive development, indices of the attachment relationship move away from
behavioral manifestations to internal representations. A few studies suggest that
it may be a viable research strategy to examine children’s working models of
attachment at the representational level.

Main et a1., ( 1985) report that the responses of 6-year-old children to a
family photograph upon physical separation from parents and answers to a
hypothetical question about a two-week separation from parents were signiﬁcantly
related to their attachment security at one year of age. They argue that children
who were securely attached in infancy developed internal representations of their
parents as being responsive to their needs and thus emotionally accessible and
that these representations were reﬂected in their responses to the attachment-
related situations. Cassidy (1988) used a puppet interview and story completion

task to access attachment representations of 6-year-old children and she found

39
qualitative differences in these childrens’ working models of the attachment

relationship. Secure children provided coherent, elaborated responses; were
ﬂexible and open in their communication style; and depicted parents as being
protective and empathic. Avoidant children actively refused to answer some
stories or responded, "I don’t know"; were non-empathic, aloof and defended in
their communication style; and presented parents as unresponsive or harsh. The
study did not identify an ambivalent pattern and the author points out that the
method was preliminary and needs to be reﬁned in further research.

Although preliminary, these studies suggest that examining child
attachment security at the representational level is feasible as children mature.
Yet, there is little research available in this area and even less with children in the
pre-school age years. Numerous researchers have commented on the need for
studying attachment and attachment-related behaviors for children in this age
range (Bretherton, Ridgeway & Cassidy, 1990; Cicchetti, Cummings, Greenberg &
Marvin, 1990; Schneider-Rosen, 1990). Therefore, the second goal of the current
study was to further investigate child attachment security and speciﬁc behavioral
parent-child interactions patterns, as a function of adult attachment security, as
they present in play interactions between mothers and their three-year-old
children.

In terms of child outcome, child attachment security was assessed with the
Attachment Story Completion Task (ASCT; Bretherton et a1., 1990), which is
designed to assess representations of attachment security in three-year-old

children. These authors note that childrens’ scores on the ASCT were associated

40
with attachment security as assessed by the Strange Situation at 18 months of age

and a concurrent measure based on separations and reunions with their mothers.
Yet, they also comment that the ASCT" is in need of further validation. As part
of the second goal, the ASCT was used to provide that additional validation. In
addition, these authors found associations among cognitive skills at 18 and 24
months of age, as assessed by the Bayley Scales of Infant Development, tests of
language development, and child attachment security at three years of age.
However, the BSID provides a developmental index that is not directly equivalent
to intelligence and the authors note that during the 24-month old assessment,
there was a signiﬁcant ceiling effect with this instrument. Thus, the current study
examined whether this relation between child attachment security and cognitive
functioning continued to be present on a standardized and comprehensive
measure of intelligence for children of pre-school age.

Finally, with regards to the third goal, no previous studies have directly
examined the variables and experiences underlying the process of reorganization
of attachment representations. It was expected that in the context of unloving
relationships with one’s parents during childhood, experiencing a loving
relationship with some other adult would be a central determinant in whether an
individual had been able to reorganize their representational model of attachment
in adulthood. The rationale for this hypothesis rests on the assumption that
having had a secure relationship with some other person would provide competing
input about issues relating to safety, security, trust, self/other representations, and

the experience of being emotionally connected to someone. This information

41

would be critical for disconﬁrming the self/other representations formed in
response to negative parent-child relationships and would provide the basis for
allowing the individual to feel safe enough to become emotionally connected in a
"corrective" relationship at a later point in life. Thus, accessing childhood pain
and engaging in a "signiﬁcant emotional experience" may be key elements in
modifying representational models, but the ability to successfully engage in this
kind of experience and process this pain would be rooted in a secure attachment
at an earlier point in the individual’s developmental history. Since the plasticity
of these representational models is not known, it was not clear whether this
experience must occur simultaneously with early parent-child relationships, or
whether it could still be effective if experienced at a later point in development.
Given that adolescence is a period of physical and psychological reorganization, it
is possible that this stage of development may pose another signiﬁcant
opportunity for disconﬁrmation of one’s representational model of attachment
However, the Crandell (1992) study suggests that if there were no loving
relationships before middle adolescence, those that occurred in later life were not
effective in altering self/other representations. Thus, it was expected that
individuals who had resolved the experience of unloving parent-child relationships
and had reorganized their representational models of attachment would have
experienced a loving relationship with some other adult either during childhood
or early adolescence while individuals who were still entangled and struggling with

these issues would not have had a secure alternate attachment relationship.

42

In addition, psychotherapy may be a signiﬁcant factor in the reorganization
of attachment representations. Many theorists and clinicians ﬁom a
psychodynamic perspective speculate that the mechanism for therapeutic change is
altering a person’s self/other representations. It is interesting to note that all of
the women ﬁom the Crandell (1992) study had been involved in psychotherapy at
some point in their lives. However, only the women who had experienced an
alternate attachment relationship demonstrated that the therapeutic experience
led to a positive change in their lives and relationships with others. These data
are only preliminary and open to a range of criticisms. However, they do indicate
that women with insecure attachment representations were involved in adult
relationships marked by intense emotional conﬂict, distrust, and dissatisfaction
and they suggest that perhaps this difﬁculty to engage successfully with others may
extend to therapeutic relationships as well. Therefore, this study also sought to
identify whether a secure alternate attachment ﬁgure during childhood was
related to experiences in psychotherapy at a later point in life and to what extent
psychotherapy was related to the reorganizational process.

Finally, the study aimed to specify to what extent intelligence may impact
the process of reorganizing representational models of attachment in adulthood.
While intelligence and attachment security in general should be independent from
each other, intelligence may be a significant factor in the. reorganization of
attachment representations given that this process involves actively accessing and
integrating experiences from one’s past into a current state of mind with respect

to attachment relationships.

43
times

Five major hypotheses were derived from the three research goals. The
ﬁrst hypothesis related to methodology, the second and third hypotheses related
to replication processes across generations, the fourth hypothesis related to the
reorganizational process of attachment representations in adulthood, and the ﬁfth
hypothesis reﬂected an exploratory question about adults with mixed

representational models of attachment. The speciﬁc hypotheses were as follows:

Hymthesis 1
It was predicted that mothers classiﬁed as secure and resolved, as assessed
by the AAQ, would report less distressed parent-child relationships than mothers

classiﬁed as insecure, as indicated by lower scores on the Parenting Stress Index

(PSI; Abidin, 1986).

Hmthesis 2

Adult attachment security would differentiate parenting behavior with
respect to the quality of maternal affect and style of relating in the following
ways:

a) Secure and resolved mothers would demonstrate more

positive affect and less negative affect when interacting

with their children than insecure mothers.

b)

Hymthesis 3

44

Secure and resolved mothers would demonstrate a more positive
style of relating when interacting with their children than insecure
mothers.

Secure and resolved mothers would demonstrate more
synchronous interactions when playing with their

children than insecure mothers.

Adult attachment security would differentiate child functioning in the

following ways:

a)

b)

Children of secure and children of resolved mothers
would demonstrate more positive and less negative
affect when interacting with their mothers than
children of insecure mothers.

Children of secure and children of resolved mothers
would demonstrate more positive and less negative
social behaviors when interacting with their mothers
than children of insecure mothers.

Children of secure and children of resolved mothers
would be classified as securely attached whereas
children of insecure mothers would be classified as

insecurely attached.

45
(1) Children of secure and children of resolved mothers

would demonstrate higher intellectual functioning than

children of insecure mothers.

W
Although resolved and insecure mothers have in common the experience of
unloving and painful parent-child relationships in childhood, they are distinctly
different in the way they have come to terms with those experiences. Therefore,
it was predicted that resolved mothers would be signiﬁcantly different from
insecure mothers in terms of their relationship histories and current functioning.
Speciﬁcally, it was predicted that:
a) Resolved mothers would have experienced a secure
alternate attachment ﬁgure in childhood or
adolescence whereas insecure mothers would not have
had this experience.
b) Resolved mothers would have more psychotherapy
experience than insecure mothers.
c) Resolved mothers would demonstrate higher

intellectual functioning than insecure mothers.

Hypothesis 5

Finally, as exploratory analyses, the sequelae for mothers with mixed

attachment representations was examined. This group was defined as those

46

mothers who demonstrated qualities of secure and insecure attachment
representations and/or experienced early parent-child relationships that were
innocuous but not actively loving. It was expected that these mothers would fall
mid-way between mothers with secure and insecure attachment representations
with regards to the quality of maternal affect and style of relating during parent-
child play interactions. It was also expected that children of mothers with mixed
attachment representations would fall mid-way between children of mothers with
secure and insecure attachment representations with respect to the quality of
child affect and social behaviors during parent-child play interactions, child
attachment security, and intellectual functioning.

The overall model of this study is illustrated in Figure 1.

47

Time

ET

 

Seouogm _
EEO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ massages“ 1’
T .8 E30
+ ﬂ ooeocomxm J A j o E
38255me _ eco§o§<
._ - m l.. 08882
l mcouﬁcomoaom _ wooAEEU E
_ Eon—€83. manages—om EEO
I :8... ﬁ seems to €90
* ooeowﬁoEL
.....-.-..-u. ..... 2.... . . . ....................a. ..............u . ...).. .....

CHAPTER TWO
Method
Subject Selegp'op

Mother-toddler dyads were recruited from birth announcements
published in a local newspaper in 1989. Twenty-two hundred women were
identiﬁed, sent an introductory letter inviting them to participate in a study of
parent-child relationships, and asked to return demographic forms to indicate
their consent. Subjects were eligible to participate if their child was not born with
any major birth defects, had no diagnosed major physical, mental or emotional
conditions, and lived in the subject’s home. Of those women who returned the
consent and demographic forms, one hundred and twenty-eight subjects met these
criteria and were subsequently mailed two questionnaires which constituted Phase
I of the study. It should be noted that the task demands of completing these
questionnaires were not insignificant. In particular, one of the questionnaires
asked personal and often painful questions about experiences in one’s family of
origin and required approximately two hours to complete.

On the basis of the two questionnaires, 70 women were invited to
continue in Phase II, of which 46 agreed and subsequently completed their
participation in the study. (Three women agreed to participate but dropped out
before completing the project.) Therefore, the total sample consisted of 46

mother-toddler dyads.

48

49

The names of all subjects who completed the project were pooled and a
lottery was drawn so that 10 subjects received a stipend of $15.00 for their

participation.

Subject Descriptiop

The mothers were predominately Caucasian (98%), married (94%), well-
educated (46% had earned a college degree and 26% had attended some college),
and in their thirties (M=33 years, §Q=4 years). Occupational status was
established using the Revised Duncan Socioeconomic Index (TSEIZ; Stevens &
Featherman, 1981). The mean occupational score for mothers was 45 (S_I2=21,
M82: 17-83), spanning the scale from janitor to university professor. The mean
occupational score for fathers was 50 (_SQ=20, BEE: 16-88), spanning the scale
from low-skilled laborer to physicist. Both mean scores represent middle-class
socioeconomic status, encompassing such occupations as nursing, teaching and
ofﬁce managing.

There were slightly more female toddlers (57%) than male (43%) and
the mean age was 40 months (S12=2 months, _R_apgg=35-46 months). The
majority of children (96%) were delivered between 36-42 gestational weeks and
78% had no birth difficulties. Of the 10 subjects who did encounter birth
difficulties, half involved an emergency Cesarean birth (primarily due to breech
presentations), with the remaining 5 subjects encountering problems involving
maternal infection or infant respiratory distress. Slightly more than half of the

toddlers (59%) attended day care, with 70% of that care being provided in a

50
private home setting and 30% in a public facility. The mean age for beginning

day care was 3 months, (§Q=9 months, an e=1~40 months). Among those
children who attended day care, 74% attended between 20-40 hours per week
while the remaining children attended less than 20 hours per week. Half of the
children in this study (52%) were the second born child in a family with two

children.

Measures and Procedure for Phase I

Women who agreed to participate were sent the Parenting Stress Index
(PSI; Abidin, 1986) and the Adult Attachment Questionnaire (AAQ; Appendix
A). The PS1 is a standardized loo-item questionnaire that assesses stress in the
parent-child relationship, providing three global ratings of parental stress: Child
Domain Score; Parent Domain Score; and Total Stress Score. The Child Domain
evaluates six dimensions of child functioning (Adaptability, Acceptability,
Demandingness, Mood, Hyperactivity, and Positive Reinforcement). The Parent
Domain evaluates seven dimensions of parent functioning (Depression,
Attachment, Competence, Social Isolation, Relationship with Spouse, Health, and
Role Restriction). These two domain scores are then summed to provide the
Total Stress Score. Abidin (1986) reports internal consistency reliability
coefficients of .89, .93 and .95 for the Child, Parent and Total Stress domains,

respectively and retest reliability coefficients of .63, .91 and .96 for these domains,

respectively.

51
The AAI was converted into a questionnaire (AAQ) by translating the

questions, with some modiﬁcations, into an 18-page, open-ended questionnaire.
Bakersman-Kranenburg & van Ijzendoorn (1993) report on the psychometric
properties of the AAI. They interviewed 83 mothers twice, two months apart with
different interviewers, and found that 78% of the AAI classiﬁcations remained
the same over this time period. In addition, they found that attachment security
classiﬁcations were unrelated to maternal intelligence, memory, or social
desirability. Since the current study was the ﬁrst one to utilize the AAQ for
classifying adult attachment security, there was no comparable information about
its reliability and validity but given that it is a derivative of the AAI, it was
reasonable to expect similar properties in identifying representational models of
attachment.

Like the AAI, the AAQ asks subjects to describe their childhood
relationships with their parents, their understanding of why their parents behaved
the way they did, the effects of these early relationships on their adult personality,
and how their relationships with their parents have changed over the years. A
modiﬁed scoring system (Adams, 1992) of the original Adult Attachment Rating
and Classification System (Main & Goldwyn, 1989) was used in the current study.
As with the original scoring system, the modified system rates responses along two
dimensions. The first dimension assesses the "emotional quality of the parent-
child relationship" and consists of the following five Experiencing Scales:
Rejection, Role-Reversal, Neglect, Pressure to Achieve; and Abuse. An overall

index of the degree to which the parent was loving toward the subject is assigned

52

on the basis of scores along these ﬁve scales. Subjects receive scores for both
parents on the Experiencing and Loving Scales. The second dimension assesses
the subject’s "current representational model of attachment" by evaluating the
structural and organizational nature of the subjects’ recollections and it consists of
the following three Representational Scales: Idealization of the parent (the
subject receives a score for each parent and an Overall Idealization score), Ability
to Recall Memories, and Resolution of Emotional Conﬂicts with the parent. An
overall index of Coherency is assigned on the basis of scores along these three
scales as well as the overall organization and connectedness of the material
presented.

One of the changes in the modiﬁed scoring system is that the rating
scales are reduced from nine-point to ﬁve-point scales. In the present study,
scores of 4 and 5 on the Experiencing and Loving Scales indicated severely
negative experiences and unloving relationships. Scores of 4 and 5 on the
Representational Scales indicated extreme defensiveness, poor ability to recall,
poor resolution and incoherent presentations. Subjects were assigned to
attachment groups on the basis of their AAQ scores by the following criteria.
Individuals were classified as resolved (Main & Goldwyn referred to this group as
”earned secure") if they had experienced a severely unloving relationship with at
least one parent and provided an integrated, coherent presentation of those
relationships (N=11 and 8/11 had severely unloving relationships with both
parents). Individuals were classiﬁed as secure if they had experienced an actively

loving relationship with at least one parent, a minimum of an adequately loving

53
relationship with the other parent, and provided an integrated, coherent

presentation of those relationships (N= 10). Individuals were classiﬁed as
insecure if they experienced a severely unloving relationship with at least one
parent and provided a highly idealized, defensive and/or incoherent presentation
of those relationships (N=15). No attempt was made to categorize individuals by
speciﬁc type of insecure attachment. There were also ten individuals who either
received Coherency scores of 3 and/or parent Loving scores of 3. These
individuals were assigned to a separate group, entitled mixed attachment
representations, for the purpose of exploratory analyses.

To assess the inter-rater reliability of the AAQ and the modiﬁed scoring
system, I trained an advanced undergraduate student in weekly 4-hour training
sessions over a three-month time period, for a total training time of
approximately 50 hours. Training consisted of scoring six practice protocols that
were thoroughly discussed to familiarize the student with the scoring system. An
additional ten training protocols were scored independently by myself and the
student for the purpose of establishing reliability. Pearson inter-rater reliability
coefficients for these ten training protocols were as follows: Rejection=.84; Role
Reversal=.86; Neglect=.81; Pressure to Achieve=.94; Abuse=.97; Loving=.87;
Idealization of Parent=.74; Recall=.69; Resolution of Conﬂict=.86;
Coherency=.90. These coefficients were slightly lower than those reported by
Adams (1992) who utilized percentage of agreement within one point to calculate

reliabilities.

54
I scored all AAQ protocols as they were returned by mail to classify

subjects by their attachment representations. Following this initial scoring, the
second coder, who was blind to my scoring and the hypotheses, scored the
protocols independently. Disagreements of one point were assigned the score
designated in the initial scoring, whereas disagreements of two or more points

were resolved through discussion between both raters.

Measures and Procedure for Phase II

Once subjects were assigned to attachment groups, they were asked to
participate in Phase II, which consisted of two sessions. In the ﬁrst session,
mothers brought their toddlers to the university campus. Mothers oriented their
child to the play room and then exited to an adjacent room across the hall. An
assistant conducted an intellectual screening with the mothers while I
administered an attachment security assessment and intellectual assessment of
their child in the playroom. Upon completion, mothers were reunited with their
child in the playroom and the dyad engaged in a 30-minute parent-child
interaction task that was videotaped behind a one-way mirror. The rationale for
this order was to provide a natural separation between parent and child to
heighten attachment themes for the child attachment assessment.

The intellectual screening used to assess maternal IQ consisted of the
Vocabulary, Arithmetic, Picture Arrangement and Block Design subtests of the
Wechsler Adult Intelligence Scale - Revised (WAlS-R; Wechsler, 1981).

Silverstein (1982) presents normative data and IQ equivalents for this

55
combination of subtests and reports a correlation of .94 between these estimated
IQ scores and Full Score IQ scores. The assistants who administered the
intellectual screening to mothers were two graduate students in the clinical
psychology program and one advanced undergraduate student. The
undergraduate student received training to administer the WAIS-R by
administering two practice screenings and eight additional screenings over a one
month period.

Child attachment security was assessed with the Attachment Story
Completion Task (ASCT; Bretherton et a1., 1990), which is designed to assess
attachment representations in three-year-old children. It consists of six family
ﬁgures (one female and one male child, two parents, and two grandparents) and a
few additional props. The child is given the beginnings to six stories, designed to
elicit attachment themes, and asked, "show me and tell me what happens next."
The child protagonist is the same sex as the subject. The ﬁrst story is a warm-up
story to orient the child to the task and is not scored. The five scored stories are
as follows:

1. Spilled Juice. The two children and parents are

seated at a table and the child figure is made to spill
his or her juice on the ﬂoor. The mother figure
responds "(George or Jane), you spilled your juice"
and the child is asked to provide an ending to the
story.

2. Hurt Knee. The two children and parents take a

pretend walk in the park. The child figure is made to
climb a tower, falling and hurting his or her knee.

56
3. mm. The child ﬁgure ascends the stairs
of the playhouse to go to bed. Upon arriving in his or
her bedroom, the child exclaims, "Mommy, daddy --
there’s a monster in my room!"
4. Mpg. The parent ﬁgures take an overnight trip,
leaving the two child ﬁgures with the grandparents.
After the parents have departed, the child is asked to
depict what the family does during their absence.
5. Reunion. The parent ﬁgures return from their trip
the following morning and the child is asked to depict
how the family reacts to their return.
These stories were videotaped and scored for attachment security using
the criteria described by the authors as guidelines. Responses were scored on a
four-point scale, with 1=very secure, 2=fairly secure, 3=insecure, and 4=very
insecure. Attachment scores were assigned based on the structure and content of
stories. Stories were scored as very secure if the child provided an appropriate
response and with little or no prompting. They were scored as fairly secure if the
child provided an appropriate response after one or two prompts and/or
demonstrated slight avoidance of the story issue. Stories were scored as insecure
if the child provided an appropriate response but only after multiple prompts,
provided an inapprOpriate response, or could not provide an ending to the story.
Stories were scored as very insecure if the child provided a bizarre ending to the
story. Examples of an appropriate response included: having the parent clean-up
the juice, providing a band-aid, alleviating the child’s fear, engaging in adaptive
behavior during separation, and reuniting the figures during reunion.

Inappropriate responses included: extreme punishment, ignoring the pain and/or

fear of the child figure, continuous searching and anxiety during separation,

57

ignoring the parent ﬁgures during reunion, and failure to provide story endings.
Examples of bizarre responses included: having the parent ﬁgure kick the child
ﬁgure in the head after falling ﬁ'om the tower, having the monster eat the child
ﬁgure, and disorganized, tangential story endings that were unrelated to the story
issue.

On the basis of these ﬁve stories, subjects were assigned an overall
attachment security score. This overall score represented the modal response
across the ﬁve stories. For example, if the child received two scores of two and
three scores of three, or one score of one, one score of four, and three scores of
three, the overall attachment security score was a three. The exception to this
rule involved scores of four. In those cases, even if the child received four scores
of one and one score of four, the overall security score was a two. For cases that
did not have a modal score, the outlying score was used to decide the direction of
the overall security score. For example, if the child received two scores of one,
two scores of two and one score of three, the overall attachment score was a two.
If there were two scores of two, two scores of three and one score of one, the
child received an overall attachment score of two.

Child IQ was assessed with a standardized, abbreviated battery of the
Stanford-Binet Intelligence Scale (SB; Thorndike, Hagen & Sattler, 1986) that
consisted of the following subtests: Vocabulary and Comprehension, Pattern
Analysis, Quantitative, Bead Memory and Memory for Sentences. Standard Area
Scores in Verbal Reasoning, Abstract/Visual Reasoning, Quantitative Reasoning,

and Short-Term Memory were obtained from these subtests, respectively. IQ

58
scores were derived from the Sum of Standard Area Scores, as indicated in the

scoring manual.

The parent-child play interaction task was comprised of three conditions:
a lO-minute child directed play session (CD); a 10minute parent directed play
session (PD); and clean-up (CU) (Robinson & Eyberg, 1981). This task was later
coded with a modiﬁed version (Whipple, Denburg & Davies, 1993) of the Belsky
Parent-Child Interaction Coding System (Belsky, Youngblade, Rovine & Volling,
1990). The revised version contains a total of 9 parent scales and 7 child scales
that are coded minute-by-minute for each 10-minute play session, and every 15
seconds during clean-up. All scales are based on a ﬁve point rating system, with
the exception of two parent scales, which are based on a three point rating
system. Scores are assigned on the basis of both frequency and intensity of verbal
and non-verbal behaviors. The scores are then summed across each scale to'
provide a total score within the CD, PD and CU conditions.

The parent scales tap two overall dimensions of warmth/affection and
control. The warmth/affection dimension is comprised of three scales. Positive
Affect assesses parental nurturance, affection, enjoyment, and enthusiasm.
Negative Affect taps parental hostility, displeasure, annoyance, sarcasm, and
escalations leading to loss of control. Parent Positive Feedback assesses the
degree to which the parent provides contingent rewards and praise to the child
for his/her behavior. The parental control dimension is comprised of four scales.
Self-Regulation assesses the parent’s ability to "be there" for the child and provide

the "scaffolding" which is supportive of the child’s efforts. Intrusive-

59

Overcontrolling taps "ill-timed" or manipulative behavior that reﬂects the parent’s
agenda, with disregard for the child’s needs. Unresponsive-Undercontrolling
captures the extent to which the parent either actively ignores the child’s request
or is uninvolved at times when support or assistance would be helpful to the child.
Demands Self-Reliant Behavior assesses parental behavior directed toward
promoting the child’s autonomous behavior without the element of facilitation
present in the Self-Regulation scale. Finally, there are two additional scales in
the parent domain. One scale assesses whether the parent Complies with the task
instructions inherent in each condition, and the other scale assesses the extent to
which the dyad engages in Parallel Play (which is not coded in the CU condition).
These two scales are based on a three point rating system.

Child affect is comprised of three scales. Positive Affect assesses the
child’s enthusiasm and level of comfort, ranging from depressed/ﬂat affect to
neutral, to expressions of fun, delight and "bubbliness." Negative Affect taps
displays of anger, dislike, hostility, resistance and aggression, whereas Degree of
Distress assesses the presence of crying, whining and expressions of fear. Child
social behavior is comprised of four scales. Dyadic Proximity encompasses both
physical and psychological closeness to the parent. Compliance involves speciﬁc
instances of complying with parental requests or instructions, whereas
Disobedience assesses overt refusals to comply. Finally, Dependency taps the
extent to which the child turns to theparent for help.

In addition, there are ﬁve global ratings completed at the end of each

condition. Global scores assess the following areas: amount of effort the parent

60
puts forth to engage the child; the degree of synchrony in the parent-child dyad;

how well the rater likes the parent; likes the child; and the child’s activity level.
Each item is scored on a ﬁve point scale, from little or none, to average, to a
great deal.

Videotapes from the MSU Longitudinal Study (Zucker & Fitzgerald,
1991) were used as training tapes for the current study. The ﬁrst author from
Whipple, Fitzgerald & Zucker (in press) trained four undergraduate students as
video coders bimonthly, over a three-month period. Each coder received a
minimum of 45 hours of training. Raters were blind to all other aspects of the
current study. Upon completion, raters independently scored a videotape
completed by the trainer. Overall Pearson correlation coefﬁcients for each rater,
compared to the trainer were as follows: .98, .96, .95, and .97. During data
collection, the four coders were divided into two dyads and each dyad was
randomly assigned 23 videotapes from attachment groups. Each pair scored
approximately every other tape so that 13 videotapes were dual coded. Inter-rater
reliability coefficients collapsing all three conditions were as follows: Parent
Scales=.98; Child Scales=.97 and Global Ratings=.63. For data entry, the dual
coded tapes were assigned the scores given by the predesignated senior coding
dyad.

In the second session, mothers returned to the university campus
without their child and completed the Attachment Relationships Interview (ARI;

Crandell, 1991; Appendix B) with the assistant who had conducted the intellectual

61

assessment in the ﬁrst session. Subjects were provided feedback about the nature
of the study following completion of this session.

All three assistants received training for the ARI by conducting three
practice interviews over a six-week period and participating in group discussions.
The ARI was used to assess the presence of an alternate attachment ﬁgure in
childhood and adolescence, and experiences of psychotherapy in adolescence and
adulthood. For descriptive analyses, alternate attachment ﬁgure was coded in six
categories, with 0=none, l=sibling, 2= grandparent, 3=aunt, 4=teacher, 5=step-
parent. For statistical analyses, alternate attachment figure was recoded as a
dichotomous variable, with 0=no and 1=yes. While all attachment relationships
are close relationships, not all close relationships represent attachment
relationships. Therefore, scoring for the ARI differentiated close and attachment
relationships by three criteria. Subjects were asked separately for childhood and
adolescence, "When you were upset, in trouble, or needing help, who did you turn
to?" They were then asked to provide detailed descriptions of those relationships,
how they evolved over time, and what they learned from them. To be scored
positive for an alternate attachment ﬁgure, the subject had to: 1) describe a
loving relationship with an elder; 2) provide concrete examples of receiving
comfort, nurturance, and emotional support in times of distress; and 3) provide
evidence that this comfort and support was a soothing experience and not simply
administrative efforts on the part of the caregiver to take care of the subject’s
distress. Descriptions of close relationships (i.e., peer relationships) were not

counted as alternate attachment ﬁgures because technically they do not adhere to

62

the essence of an attachment relationship that is based on caregiving experiences
from an individual who is in a position to provide protection and security.

On the ARI, subjects were also asked whether they had any
psychotherapy in adolescence or adulthood, and they were asked to provide
descriptions of the type and duration of this therapy. For descriptive analyses,
type of psychotherapy was coded in ﬁve categories, with 0=none, 1=individual,
2=marital, 3=family, and 4: group. For statistical analyses, this variable was
recoded so that 0=no psychotherapy experience and 1=positive for psychotherapy
experience. Duration was recoded in ﬁve categories with 0=none, l=less than
three months, 2=three to six months, 3=six+ to twelve months, 4=twelve+ to
twenty-four months, and 5=more than twenty-four months. Scores were based on
continuous psychotherapy so that an eleventh-month period in adolescence and a
twelve-month period in adulthood would receive scores of 3 for each time period,
rather than a score of 4 for adulthood.

A summary of the measures used is presented in Table 1.

63
Table 1

Summary of Measures

 

 

MEASURES

PHASE I

Mother
Historical Information and Adult Attachment Questionnaire
Adult Attachment Security
Self-Perception of Distress in Parenting Stress Index
the Parent-Child Relationship

PHASE 11

Mother
Intelligence WAIS-R (abbreviated form)
Alternate Attachment Figure Attachment Relationships Interview
and Psychotherapy Experience

Child
Intelligence Stanford-Binet (abbreviated form)
Attachment Security Attachment Story Completion Task

Parent-Child mad

Parent-Child Interaction Belsky Parent-Child Interaction
Coding System

 

CHAPTER THREE
Results

Overall, the data supported the predicted model illustrated in Figure 1.
Results are discussed in ﬁve sections, corresponding to the speciﬁc hypotheses
detailed in Chapter One, but prior to the presentation of this data, the
psychometric properties of the AAQ are reported.

Table 2 presents the inter-rater reliabilities for the individual and overall
scales of the AAQ. The correlations ranged from .78 to .96 on the separate
scales. The reliabilities for the Loving and Coherent scales, which were used to
classify individuals with respect to their attachment security, were .81 and .78,
respectively. As indicated in Table 3, AAQ classifications were not completely
independent of intellectual functioning. Mothers classified as secure scored
signiﬁcantly higher than mothers classiﬁed as insecure on Vocabulary and they
scored higher than insecure and mixed mothers on overall IQ. Otherwise, there
were no signiﬁcant differences among any of the four attachment groups on any

of the subtests scores or overall IQ scores.

Hypothesis 1

Adult attachment security did not differentiate distressed from non-
distressed parent-child relationships on the Parenting Stress Index (PSI). As
indicated in Table 4, there were no significant differences among any of the

groups on any of the scales on the PS1. The means reported reﬂect percentile

65
Table 2

Inter-rater Reliabilities for the Adult Attachment uestionnaire

 

Pearson Correlations

gmriencing Scales

Rejection .85
Role Reversal .89
Neglect .96
Pressure to Achieve .89
Abuse .95
Loving .81

Representational Scales

Idealization .87
Degree of Recall .83
Resolution .78
Coherency .78
Average of all Scales Combined .87

 

Correlations are based on the scores of two raters

reﬁners
2:15 can: ﬁne 2885 815 seem 513829...”—
.ANme amoumhozmm so: ﬂ0>mhov mouOOm Cm @8979:— .mo.~oom 2&8 ﬁuaﬁﬂgw mum—<3 goo—mo.— 3.302”

m—ﬂOV—uﬂagoziaﬁvﬁaum "MZW—

 

66

 

 

 

 

 

 

33 Am :5. mod 3 .b 8.3— 3.3 $.33 Haw 3.x 2 mos 3.63 O—
3. 3d —~.N 8. 3 $.N 8.3 cod 8.3 3.4 we." _ .moQ goo—m
3. 5.3 3.3 3.3 om.N cad 3.3 em. 3 8.~ 3d .t< .63
mm. a; 8.3 8.3 a5... 3.3 Gd am. 2 we.“ 3.3 0:08:33.
Tum mc. owﬁ Sam 83 3 and 3.3 RN 3.3 Nm.~ 3.3 «b3380>
«Mlﬂw. 02.3 m olulaﬂ ﬁlm :32 NW .802 NW. :32 dim. emu—2
Qm—XHE mm 8mm; mm 30mm Qm>40mmm

 

 

67
ranks. Scores between 15th and 80th percentiles are considered to be within the

normal range. All of the means for all three groups were within this non-clinical
range. In addition, the standard deviations for all three groups were typically 20-
30 percentile points, indicating that some of the mothers in each group were
experiencing clinical levels of stress but these cases could not identified as a

function of maternal attachment security.

Hmtheses 2 and 3

Hypotheses 2 and 3 predicted that adult attachment security would
differentiate parenting and child behavior with respect to the quality of affect,
style of relating, and parent-child synchrony during play interactions. Prior to
conducting these comparisons, preliminary analyses were conducted to ascertain
whether resolved and secure mothers could be combined into a single group for
comparisons with insecure mothers. The results of these analyses are presented
in Table 5 through Table 7.

As Table 5 indicates, there were no significant differences between
resolved and secure mothers on any of the parent demographic variables.
However, there was a significant difference on one child demographic variable.
Ten of the eleven resolved mothers (91%) had a female child, compared to five
of the ten secure mothers (50%). Otherwise, there were no differences between

these groups with respect to child age, birth order, or day care experience.

68

oJI: a> M

3m.
8.—
3.3
3.3

3.
33% r.—

 

8.3
3.8
36m
mean
3.3
36m
amém
8.3a
3.3
3.3
8.8
3.3”
8.8
3.3
3.3
8.3

mm

ﬁns mmaomma

whnm
8.8
3.3
8.39
whmm
8.3.
36w
8.9m
3.3
3.3.
3.3
3.3.
wwém
8.8
8.3
3.3

:32

 

3.3
35.8
8.3m
coda
when
3.3
2.6a
3.8
aim
8.8
318
Sam
8.8
5.8
3.3
3.3

mm

8.:
86m
8.?
8.3
8.3
and.”
86m
8.3.
8.3
8.3
8.3
8.3V
8.3
8.3
33%
8.5.

:32

 

§u5 mmaumm

8.3
8.2
Emma
vmﬁm
3.3
8.3“
8.8
8.3
mwdm
chm
3.3
3.8
3.3
3.3
8.8
8.3
mm.

SE: 5508:

3.3
8.3
8.3
3.5.
made
3.3
8.8
ems
8.8
8.3.
8.8
36m
3.3.
8.8
8.3.
3.8

:32

 

92 no.3
5.33
E53 3.83m
core—83 Eoom
3:809:00
83581 23—
825.032
aomwmoaoo
33
3388303
0382095335
805.502
“ﬂaws—Essen
§=ﬁao8<
55252
5.350 EEO
mmmmhm 173.03.

 

 

 

Eon—€32 0.508:— 15 9.30m

 

dozoam ma? meow—82 “0 meat“ :50 5...: 82% £255.—

 

v 03am.

69

Table 6 displays the results of comparisons between resolved and secure
mothers on the PSI. There were no signiﬁcant differences between these two
groups on any of the overall stress or child variables. However, there were two
interesting differences related to parental functioning. Resolved mothers scored
signiﬁcantly higher on Spouse Relationship than secure mothers, indicating that
they perceived more emotional distance and lack of support from their spouses
regarding child management. In addition, there was a non-significant trend on
the Depression scale, suggesting that resolved mothers experienced more sadness,
guilt, and/or dissatisfaction with themselves and their life circumstances than
secure mothers.

Table 7 presents the comparisons of resolved and secure mothers with
respect to parent-child interactions. There were no significant differences
between these two groups on any of the parenting scales of the parent-child
interaction task (PCI). Additionally, there were no significant differences
between children of resolved and children of insecure mothers on any of the child
scales of the PCI.

Given that there were no significant differences between resolved and
secure mothers, or children of resolved and children of secure mothers with
regard to parent-child play interactions, these two groups were joined to form a
combined secure group (n=21) for comparisons with insecure mothers (n=15) to
test hypotheses 2 and 3. As indicated in Table 8, the results from these analyses
supported the prediction that adult attachment security would differentiate

parenting behavior with respect to the quality of parent affect, style of relating,

70

 

 

Table 5
De 'c Va 'abl of F ' 'es with eso ed a (1 Secure Mate ttac e t
RESOLVED (n2 11) SECURE (n: 10)
M S_D M_e_au £12 PM mm
Mother Age 35.37 3.92 35.59 3.73 .01 .90
(in years)
Mother Educ.1 4.18 .60 4.30 .68 .18 .68
Mother Race 3.00 .00 3.00 .00 ~--- «-
(3=Caueasian)
Marital Status 3.09 .30 3.00 .00 .90 .35
(3=Mar 4=Sin)
Mother Occup.2 46.76 22.69 59.74 19.67 1.95 .18
Father Occup.2 55.32 21.38 60.56 12.94 .45 .51
Mother # kids 1.82 .60 1.80 .63 .01 .95
Blended FamilyJ 1.91 .30 1.90 .32 .01 .95
Length of Preg.‘ 1.92 .30 2.00 .00 .91 .35
Birth Diff.’ 1.64 .51 1.80 .42 .64 .43
Birth Order 1.55 .69 1.70 .48 .35 .56
Child Sex 1.09 .30 1.50 .53 4.89 .04
(1=F 2=M)
Child Age 40.09 2.81 40.00 3.16 .01 .95
(in months)
DAY CARE
Attends3 1.36 .50 1.20 .42 .64 .43
Type‘ 1.43 .54 1.88 .35 3.74 .08
Start (in mon) 13.29 14.87 7.75 9.11 .78 .39
Frequency‘5 1.57 .54 1.88 .35 1.73 .21

 

1Mother Educ: 4=college degree and 5 = graduate degree

2Occupation scores derived from Duncan Socioeconomic Index (Stevens & Featherman, 1981)

3Dichotomous variable: 1=yes and 2=no

‘Length of pregnancy: 1= <36 weeks, 2:36-42 weeks, 3= >42 weeks

’T‘ype of daycare: 1=pub1ic facility, 2=home care, 3=other
“Frequency: 1= <20 hrs/wk, 2=20-40 hr/wk, 3= >40 hr/Wk
(g1: 1,19)

71

 

 

 

RESOLVED SECURE
(n: 11) (n2 10)
M91! S_D ean SQ E 2
Ratio Value
Total Stress 58.91 27.07 47.00 31.90 .86 .37
Child Domain 6.64 1.36 6.80 .92 .10 .75
Adaptability 63.00 25.43 65.00 28.87 .02 .87
Aoceptability 42.64 31.71 54.50 33.70 .69 .41
Demandingness 56.18 28.95 49.80 35.67 .20 .66
Moodiness 58.09 32.53 55.90 28.49 .03 .87
Dist/Hyper 44.09 34.56 51.50 36.82 .23 .64
Reinforcing 50.82 33.83 53.50 24.39 .04 .84
Parent Domain 6.46 1.37 6.60 .97 .08 .78
Depression 60.09 30.30 35.00 27.79 3.88 .06
Attachment 54.91 34.80 32.50 25.19 2.80 .11
Role Restrict 62.27 17.37 54.00 26.75 .72 .41
Competence 47.27 25.24 31.60 28.66 1.78 .20
Social Iso 64.09 28.36 45.00 29.06 2.32 .14
Spouse R’ship 73.09 15.85 53.00 26.79 4.48 .05
Health 63.18 29.00 47.50 29.93 1.49 .24

 

(g_r=1,19)

._o>o_ mc.vQ 2: .a E8E:wﬁ-=o= 203 a =<

82.1.3

882.” ca 3 3:11; 53 A038 «Eon m a no com—3 2a 533 .QE 3:23 can Suﬁ—9:00 5.. 3083 28» 33a m a .5 v8.3 03 2:32
871$ 058m 23 ASHE oozomom

 

72

 

 

 

 

 

 

 

 

DU 3 £0.50 .0: 28m cc. cc. cc; cc. cc; cc. cc. cc." cc. cc; .8:— 3:23
mm; mm. whm ac. bad cc. cc. cc.m cc. cc.m SA ac. bad cc. cc.m 853388
8. ac. ac." 2. ac; I. 3. cu." 3.. mm.“ 3V. 2. a: cm. mm.“ Anzac—om
:4 vc. SA cc. cc; ca. cc. cc; mc. 84 8. mc. SA mc. S.— £8333
Er. :. mca cc. NcA wcd cc. cc; 3. SA cw; cc. cc; NH. 34 023::—
NN. a“. wad Hm. cm.m mc. mm. and cm. end cad mm. Rem am. Sam coma—om
cm. Nm. mm.~ S. VNN 5. cv. chm hm. cm.~ mm. cv. cﬁa am. Saw 6er .mom
"a. cc. ccA vc. SA :4 mc. 54 cc. cc.— cc. cc. ca. cc. ca. 83.2 .32
«c. cm. 8." cm. mm.~ cc. mv. mc.m 3. mcd 3. mm. £8 3”. mac 6&2 .8.—
Emma}—

ozam 23m 0:3—

m a g Ola :32 H mm 532 a :32 w. a 532 mm: :32
230m 3283— oSoom 3283a 230m 3283—
a: 245.5 ><Am Qmﬂbmm—D Hzmm<m >5; DEmEQ 94—20

 

 

“nos—.0334 35032 23% one @0283— .. 33:.»— :26225 220-295.—

h 2.3

73

.05. mc.va 0.: “a Exoamcwmméoa 08 823—. n— =<

sing

ca 3 gonna 5?» .208 «ﬁg m a no V83 03 8802
A3 "5 0500.0. can ASHE c0280.:

 

 

 

~c.— va. 8.». ac. mad Nc. vw. Sam Hm. cad me; an. cN.m mm. nm.m
S. «v. cud cm. ac.m cc. 5.. cc.m mc. cc.m c_. an. cad cc. Sc
vc. an. 9mm ac. cvsm 2. ch. cud an. 55 ac. mm. calm ac. cm.m
8. cc. cvd ac. cm.m am. mm. chm an. bad cc. an. cﬁm aw. cﬁm
2. EA cﬂm 5c. vcd mc. cc. chm ac. vcd a. as. cc.~ em. ac.m
caA cc. cc; nc. «ca 8. ac. mcé wc. mca mc.~ mc. SA ac. 3.—
84 cm. EA mc. ac.“ Ma. cc. cc; mc. ac; _c. mc. 84 ac. SA
an.“ 34 Ram cc. mud. and mm. 3.." an. cm; cc. cc. SA cu. c:
av. cm. mud cm. and no. we. win an. cud ~a.N cm. and mm. and
a“; ch. RA mc. «c; Na. 3. ac; vc. Nc.— cc. cc. cc; cc. cc;
nm.~ 3.. cc." cc. cc.“ an; 3. SA cc. cc." :4 mc. 84 cc. cc;
cc. Nm. ch Na. ~a.~ ac. cm. ac.m 3. Ed mc. nc. vcd cm. cc.m
exam 0:3— 28m
M dim: mg. .qm. .502 m. mm. 802 mm :32 M NW .302 .Dlm .802
0300c V0280“ 0300a 3283— 0500c c0280”—
59 77%—AU >85 OMB—Oman Hzmmaﬁ ><Am DEUm—ED QA—IU

 

 

 

 

 

 

 

5% o-..
2&2 220
00:4 220
00:.— E085—
0waw=m

mémOu—O
3:02.300

6385
02.33800
88:28
80:85
.882 .82
.082 .8;
a

20.0.53 a maﬁa.

74

and degree of synchrony in parent-child play interactions. The results also
supported the prediction that adult attachment security would differentiate child
functioning with regard to the quality of affect and social behavior during parent-
child play interactions.

As illustrated in Table 8, in the child-directed play condition (CD) of
the PCI, there were no differences between the combined secure and insecure
mothers on any of the parent scales. However, children of insecure mothers had
significantly higher levels of negative affect and distress than children of combined
secure mothers. In addition, on the global ratings, children of combined secure
mothers were rated as more likeable and these parent-child dyads engaged in
more synchronous interactions than the insecure dyads.

Continuing with Table 8, in the parent-directed play condition (PD),
there were numerous differences between the combined secure and insecure
mothers. The insecure mothers expressed more negative affect, were more
intrusive and overcontrolling in their play, and were more likely to engage in
emotionally disconnected, parallel play than combined secure mothers. Children
of insecure mothers were more disobedient than children of secure mothers. In
addition, these dyads engaged in less synchronous interactions than the combined
secure dyads. There was also a non-significant trend, suggesting that combined
secure mothers expressed more positive affect and provided more positive

feedback than insecure mothers.

80!... .52.. .39.... .29... 8:813
880...
:: .: 8:". 5.3 0.08 .50.. m a :0 .08.... 0.: :02? .3... 0:25. .6: 00::.....:00 E02... .0. 2.00:0. 0.8 3.0.. m a :0 080.. 8:00:

3.":v 0508:. .5: CNHS 0.:00c 60:...800

 

75

30 :. v0.50 .0: 0.00m .N... 3. cc. cc. cc. cc. cc. cc. cc. cc. am... .0..:.:..

cc. cc. 2N cm. th N... cN. aa.N cc. cc.m 2.. cc. cc.” ac. aa.N 008.3800
.ncd vc. Nc. c.. cc. cc.” nN. NN. hN. cc. cc.N N.. 3. a.. NN. ..0~...:0m
.5. cc. cc. Nc. .c. .c. vc. .c. No. .c. av. cc. cc. mc. .c. .:00:0.0:D
hm. MN. 8. cc. cc. .mNﬁ vc. c... c.. cc. Nm. c.. cc. ac. Nc. 0335:.
NW. N mm. N..m cN. cN.m ch cN. c...” cN. cm.” a. 3. mN.m 5N. Sum .c0~...:0m
uNn... MN. cc.N cN. cN.N .23 mm. .m.N mm. mnN Nc. aN. .m.N um. cv.N .000". .80..

.a. 5.. cc. mc. .c. unﬁc c.. S. Nc. .c. N... cN. 8. cc. cc. .00.? .w07.
.23 .c. mc.N mc. cc.N .acd 3.. a...N cm. ccd MN. cm. cc.N mm. a...N .082 .80..

 

 

 

 

 

 

 

 

 

...Zm....<..
0...... 0.3... 0......
H... a :00: mm :82 m. mm :82 Qlc :33. m... NW :82 mm :«02
0508:. .00c 0800 0308:. .00c ....:00 0508:. .00c 0800
m... z<m...0 ><... DEmEQ gums. ><... memmﬁ. Q....0

 

 

c 03:...

...... .62.... a. 5.... Ba .08 02...... .... 8.: 8 2.22... .8... on ... 3.8 20F :0.
..a ... 8:". ....B 0.08 8.0.. n a :0 08.... 8:002 .cc.v..u .....v..u .nc.v.... .c..v... ...M. Hg 3. "5 0508:. 0:0 ..NHE 0800c 00:...800

 

cc. c..c. ..cM. ...... Nn... .08.... 30
.c..n Mc. cc.N Mc. VNM 0.0.N. cm. ch ..c. cM.M ..M...c. 0c. SN ..c. cM.M ..08.c 0-..
..c.c N... ..c.N ...M. ...M cc. cc. .N.M mm. cc.M c0. cc. cc.M hm. cc.N 0>..0<
c..N cc. ..c.M cc. NYM nc. a... 8M M... MM.M .86 0c. th cc. cM.M 00... 0....0
NYM 0.. SN 0... cM.M cmN cc. NQN ..c. cM.M cM.M cc. cc.N M... ...M 00... 80.0..

.c. c... hN.M Ma. SM Nv. cc. ch cc. th cc. Mc. M...N ..c. 3N 0c0c8m.

c.<m.0.0

.c. Mc. .c. c. .c. cc. Mc. Mc. cc. 8.. mM. cc. Nc. ..c. Mc. 3:005:09

cc. cc. 3.. .N. ..c. ....c.m cc. Mc. NN. 8.. cc. Mc. 8. Mc. .c. 0008.0
.cn... c0. c..M cc. Nc... cc. cN. cM. cN. cM. mm. M.. M... c.. .... 008....800

c.. h... MN.M cw. .M.M Nc. ..M. ...M N... .....M cM. MN. .....M be. cm.M .3808...

cc. NN. n... Nn. v... M... 0.. cc. 0c. Nc. ..th ... cc. cc. cc. 80.8.0

Nc. c.. cc. 0N. cc. c... ... ..c. cc. Nc. .vc.m cc. 3. Nc. cc. 80.2 .c0Z
Nc. cN. cc.N ...N. ..c.N .... 0.. ..cM M. ...M .c. .N. ch c.. mc.M .00.? .80..

 

 

 

 

 

 

 

 

 

 

9.50
0...... 0...... 0.23.
...... a 8.02 mm .802 .N.. mm :002 qcu :002 m ﬁlm :002 dlc :002
0808... .03 ...800 0808:. .00m ...800 0808:. .00c ......00
..D z<m..0 >5. Gaman ...zm.~.<.. >5. Qm...0m.~..n. c..:0

90.0.80. c 0.2.0...

77

During the Clean-Up condition (CU), combined secure mothers
provided more positive feedback and encouraged more self-reliant behavior from
their children than insecure mothers. In turn, their children were more active and
compliant than children of insecure mothers. Despite the fact that combined
secure mothers directed their children toward more self-reliant behavior than
insecure mothers, these dyads were once again rated as more synchronous in their
interactions than insecure parent-child dyads. There was also a non-significant
trend, suggesting that combined secure mothers demonstrated more positive affect
than insecure mothers.

With regard to child attachment security, there was partial support for
the prediction that children of resolved and children of secure mothers would
demonstrate secure attachment whereas children of insecure mothers would
demonstrate insecure attachment. Prior to examining these differences, Table 9
displays the correlational matrix for the individual stories of the ASCT and child
intellectual functioning. All five stories were significantly correlated to each other
and to the overall security score, validating the internal consistency of the ASCT
in tapping representational models of child attachment security. Consistent with
the data reported by Bretherton, et a1., (1990) child attachment security was
related to intellectual functioning as three of the five stories and the overall

attachment score were related to Verbal Reasoning, Memory, and IQ.

c..-NZ

 

78

8.x... .82..
...-c. 8.80..
.ch. .0... :0..0.0..0c
5c. 050. .3. .3852
.8. .3. ...m. .8. 35. E...
....0. ...m. .8. ...... ....m. 83.
.N...- .- 00.. 00.- 8.- .N...- 0.
......- 8. m..- 0..- .9... 8- .8. 025.2
8.- ...- S.- o..- .0... N..- .8 a... .550
.8.- m..- 8.- 8.- mm.- .8. .... .8. m0. 08:8...
.9... vm: .....- .9..- ...? .N...- .00. .3. .0... .00. .85>

0.00m 08:02 00:0. 00....-
..00..< 8.8.0”. :0..0..0..0c 800.00.. ...... 00.....c 0. 80802 8000 .0083...

 

 

0.80... :0..0. ...00 9c 808.00.... 0:0 5.8008... 00.0208. 0....0 :08< 8000.080808.

c 0.00...

79
Table 10 presents the ANOVA for children of resolved, secure and'

insecure mothers with respect to child attachment security. The Sequence
variable refers to the order in which children were administered the ASCT. The
intent was to present the ASCT in the ﬁrst segment of the three-part assessment,
followed by the intellectual assessment and then the parent-child play interaction
task. This sequence was followed in the majority of cases. However, some
children were too anxious to separate from their mothers initially so the
assessment was reordered such that the interaction task came first, followed by
the ASCT and then the intellectual assessment. In addition to this reordering,
some children remained anxious after the parent-child interaction task and
required that their mothers remain present for the ASCT and Stanford-Binet
assessments. As illustrated in Table 10, the ANOVA indicate that there were no
signiﬁcant differences among the three attachment groups with respect to the
sequencing of the assessment or maternal presence. Descriptive data provided a
different picture. The majority of children in all attachment groups were able to
progress through the assessment sequence as planned (resolved=64%,
securc=80%, insccure=73%, and mixed=60%). However, for those children who
were not, the only children who required the presence of their mothers for the
remainder of the assessment belonged to mothers with insecure attachment.
None of the children from the mothers with resolved or secure attachment
required the presence of their mothers once they were acclimated to the

playroom.

A00.0u0.00

0:0—500:0 00:000:_uv.m 0:0 808500000 05000HN; FEB .2000 0:80. 0 0 :o .0000: 00:90 H09...
09?; 0:0 o:uo 5:3 HUm< mo 530502360 mESo 0:000:— 00? .052: 00:00:3u0:000:m 005020
808000000 220 00090005 :_ H.090. ...o 00EOH00:0:~00m0

0.:0M-:0§0Z-0:0::0mu¥2m_

 

80

32:03 0:08—60:00.

 

 

 

 

:v: 00. 00.0 00. 00.0 00. 00.0 00. 00.0 00.26
0... 00. 00. 00.0 02 00.0 00.0 00.0 02.500
00. 00.0 00. 00.0 00. 00.0 00. 00.0 8080000
00. 0:. 00. 00.0 02 00.0 00.: 00.0 .3822
V0 :v: :00. 00.0 00. 00.0 00.0 00.0 00. 00.0 85: :0:
_ .mv: :00. 00.0 00. 00.0 00. 00.0 00. 00.: .80: 02000
00. 00.: 00. 00. 00. 00. 00. 00. 05.2.: .2002
00. 00. 00. 00.: 0... 00.: :0. 00.0 .850000

.Ixuzlm 83> 0 000: 0 mm 0% mm :82 0W 532

AmHHE 0.50005 SHHE 0500m ASHE o0>_o00m

 

 

3 030B

81

Continuing with Table 10, there were significant differences among the
ASCT stories and overall child attachment security as a function of maternal
attachment security. Children of resolved mothers were scored as more secure
than children of secure and children of insecure mothers on the Spilled Juice
story. In addition, children of resolved and children of secure mothers were
scored as more secure than children of insecure mothers on the Hurt Knee story.
Finally, children of resolved mothers were scored as more secure than children of
insecure mothers with respect to their overall attachment security.

In keeping with the analyses for the parent-child interaction task,
children of resolved mothers were combined with children of secure mothers for
comparisons with children of insecure mothers. These comparisons are presented
in Table 11 and the results indicated a similar pattern. Children of the combined
secure mothers were more secure than children of insecure mothers on the
Spilled Juice and Hurt Knee stories, as well as on their overall attachment
security.

Finally, with regard to child intellectual functioning, the results
supported the prediction that children of resolved and children of secure mothers
would demonstrate higher intellectual functioning than children of insecure
mothers. As indicated in Table 12, both groups scored higher than children of
insecure mothers on Verbal Reasoning, Abstracting Reasoning, Memory, and
overall IQ. In addition, children of resolved mothers scored significantly higher

than children of secure mothers on Verbal Reasoning.

82

 

 

 

Table 11
Child Mment Securig by Combined Secure and Insecure Maternal Attachment
Combined Secure Insecure
(n=21) (n=15)
Mean SQ Mean S_D F Ratio P Value
Sequencel 1.29 .46 1.33 .62 .07 .79
Present2 .00 .00 .20 .56 2.71 .11
Spilled Juice3 2.10 1.00 2.87 .52 7.52 .01
Hurt Knee 2.14 .85 3.20 .68 15.85 .000
Monster 2.52 1.08 2.53 .99 .00 .98
Separation 2.43 .87 2.87 .83 2.30 .14
Reunion 2.33 1.07 2.73 .59 1.73 .20
Total Security 2.38 .81 2.93 .46 5.71 .03

 

1Sequence=0rdcr of ASCT in three-part child assessment

2Mother Present=whether mother was present during administration of ASCT, with 0=no
and 1=yes

3ASCI‘ stories based on a 4 point scale, with 1,2=sccure attachment and 3,4=insecurc
attachment

(gf=1,34)

83

:00.v:. .:0.v.:. .00.v0.

00.03.00

.35—38 ﬂaw—8a 2: 89¢ vogue—v any—8m 0— 6w HO Om “as can no I2: .0 £33 amp—Cow «we Ema—unwam QummOQEOU goammt—qu-«aauw «00:0.— nanny—20
m—aovméuagozuaaomvzuw " MZW:

 

 

 

 

.000: 00:: 00.00 70 00.0 000:: :A: 00.0 000:: o:
.000 00.0: 00.00 :0 00:: 000:: :A: 00.0: 00.00: 08:52
00.0 00.0: 00.00: 00.0: 00.::: 00.0: 000:: .0830
.000 00:: 00.00 :0 :00 00.::: :A: 00.0 000:: .808...-
000.0: 00.0: 00.00 70 00.0: 0.0:: :00: 00.0 0:.00: 38>
0'05

m @0- 08: 0'20 dd :82 .Ixz..0.. 0:10. :82

 

 

871.: 003000 Ens 0:0-5:000:

 

 

 

my 030%

84

In addition to these differences in parenting and child functioning, there
were also signiﬁcant differences on demographic variables between combined
secure and insecure mothers. As illustrated in Table 13, combined secure
mothers were signiﬁcantly older, more educated, and married to husbands with
higher occupational levels than insecure mothers. Otherwise, there were no
signiﬁcant differences between combined secure and insecure mothers on any
other parent or child demographic variables. (The reader is referred back to
Chapter Two for comparisons of these variables to the overall sample

characteristics.)

Supplemental Analyses

Summarizing the ANOVA’s with regard to parenting behavior, mothers with
secure and resolved attachment representations expressed more positive affect,
demonstrated more positive parenting behaviors, and engaged in more
synchronous interactions with their children than mothers with insecure
attachment representations. However, since the resolved and secure mothers
were significantly older, more educated, and of higher socioeconomic status than
insecure mothers, one could argue that these demographic variables were the
driving force underlying the different interactional patterns.

To test this latter hypothesis, a series of post hoc hierarchial multiple
regressions were conducted. The significant demographic variables were
regressed onto the PCI scales in step one and maternal attachment security in

step two. The significant results are presented in Table 14.

85

 

 

 

 

Table 13
t 'ffere in Demo a 'c Va 'ablas or Mot ers 't Co bin Secu ea d
Insecure Attachment
Combined Secure Insecure
Mean S_D Mean SD E
Maternal Age 35.47 3.74 30.41 3.75 16.06"
(in years)
Mother 4.24 .63 3.13 .83 20.70b
Education1
Father 57.81 17.64 35.65 15.10 15.53'I
Occupation2

 

1Education coded so that 3=some college, 4=oollege degree, and 5= graduate degree

2Occupation scores based on the Duncan Socioeconomic Index (Stevens & Featherman, 1981)

-p<.01, bp<.001

(101,33)

86

As indicated in Table 14, maternal attachment security accounted for a
signiﬁcant portion of the variance above and beyond the demographic variables
on a particular triad of parent-child behaviors involving parental expectations for
child autonomous behavior, child compliance, and parent-child synchrony.
Speciﬁcally, in the PD and CU conditions, maternal attachment security was
uniquely related to parental encouragement of child self-regulatory and self-
reliant behavior, and child compliance. With regard to parent-child synchrony,
maternal attachment security accounted for an additional 7%, 12% and 12% of
the variance during CD, PD, and CU conditions, respectively. These results
suggested that adult attachment security was highly correlated with maternal
education, socioeconomic status, and age at child bearing. While none of the
demographic variables exerted a significant inﬂuence independently, collectively,
they impacted the quality of parent-child play interactions. During interactions
involving parent-directed tasks (i.e., PD play and CU conditions), maternal
attachment security imparted an additional inﬂuence on parental expectations of
child behavior and child compliance. In addition, maternal attachment security
was uniquely related to the degree of parent-child synchrony in interactions

involving free play and goal-directed activities.

Hypothesis 4

The results supported the prediction that resolved mothers would be
differentiated from insecure mothers with respect to their relationships histories.

The Frequency and ANOVA results presented in Table 15 provide a descriptive

 

87

 

 

E Adj. F Value F Changg Egg I
cum
Demo .23 .17 3.96”
Age .33 1.47
Education .13 -.54
SES .14 .93
Attach Sec. .30 .23 4.34” 4.46‘ -.34 -2.11‘
W
Demo .03 -.04 .37
Age -.24 -.93
Education -.07 -.25
SES -.10 -.63
Attach Sec .11 .02 1.24 3.78‘ -.36 -1.94'
PD CHILD COMPL
Demo .06 -.01 .82
Age -.22 -.87
Education -.14 -.55
SES -.19 -1.16
Attach Sec .15 .07 1.81 4.58' -.39 -2.14'
PD SYNCH
Demo .11 .05 1.70
Age .03 .12
Education .09 .35
SES -.02 -.12
Attach Sec .23 .15 2.97‘ 6.15” -.42 -2.48”
CU SELF-REG
Demo .17 .11 2.91'
Age .27 1.14
Education -.36 -1.44
SES .28 1.84
Attach Sec .25 .18 3.37” 4.08‘ -.34 -2.02'
CU CHILD COMPL
Demo .01 -.06 .11
Age .03 .11
Education -.37 -1.48
SES -.15 -.98
Attach Sec .25 .18 3.49” 13.60c -.62 -3.69c
CU SYNCH
Demo .03 -.04 .45
Age .09 .35
Education -.10 -.38
SES -.09 -.55
Attach Sec .15 .06 1.76 5.53" -.42 -2.35‘
N =46

‘p<.05, l"p<.01, ‘p<.001

88

 

 

 

 

 

Table 15
vtam t : 1'"! i ' :udP got. ra 0.3-0.1. Resolvedand 4;" re daternal
5.1mm
RESOLVED (n=11) INSECURE (n=15)

Frgueng AN OVA Fregueng AN OVA

Desc. Mean S_D Desc. Meap S_D E
AChild‘ 64% .64 .51 13% .13 .35 9.01”
ATeen1 55% .55 .52 7% .07 .26 9.54”
AEver1 82% 13%
TI‘her2 46% .46 .52 7% .07 .26 6.26”
TDur’ 12-24 mon 1.70 2.28 >24 mon .33 1.33 3.92‘
ATher‘ 64% 60%
ADur3 >24 mon 4.14 1.87 3-6 mon 1.67 .71 13.60”
l'l‘herz 64% .64 .51 33% .33 .49 2.38
IDur3 >24 mon 4.14 1.87 3-6 mon 1.80 .84 6.77”

 

‘AChild, ATeen and AEver = presence of a secure alternate attachment ﬁgure in childhood,
adolescence, and either of those developmental stages, respectively, with 1=yes and 0=no.

2'I'I'her and ITher = individual psychotherapy in adolescence and adulthood, respectively, with
1=yes and 0=no.

3’I'Dur, ADur and IDur = months of psychotherapy
‘ATher includes individual, marital, family or group psychotherapy in adulthood

'p<.05, ”p<.01

89

account of the differences between resolved and insecure mothers with respect to
alternate attachment ﬁgures in childhood and adolescence, and psychotherapy
experiences. With regard to a secure alternate attachment ﬁgure (AAF), 64% of
the resolved mothers had an AAF in childhood compared to only 13% of the
insecure mothers and 55% of the resolved mothers had an AAF in adolescence
compared to only 7% of the insecure mothers. Collectively, nine of the eleven
women (82%) in the resolved group had an AAF at some point in their
developmental history, compared to only two of the fifteen women (13%) from
the insecure group. These differences were striking and highly significant. With
regards to psychotherapy, nearly half of the resolved mothers (46%) entered into
individual psychotherapy as an adolescent, compared to only one of the insecure
mothers (7%). More than half of each group engaged in some type of
pwchotherapy in adulthood. The differences were the type and duration of
therapy. All of the resolved mothers had individual psychotherapy, with a mean
duration of more than 24 months. In comparison, only 33% of the insecure
mothers had individual psychotherapy, with a mean duration of 3 to 6 months.
The remaining insecure mothers who engaged in therapy participated in either
marital, family or group psychotherapy, also with a mean duration of 3 to 6
months. Thus, resolved mothers were more likely to enter into individual
psychotherapy in adulthood than insecure mothers and they stayed engaged in
therapy for a significantly longer time.

While the number of subjects with no AAF in their histories was too small

to conduct an ANOVA, descriptive data suggested that even more intensive

90

individual psychotherapy was related to the reorganization of attachment
representations for these individuals. Regarding the two resolved subjects who
never had an AAF, one mother participated in individual psychotherapy in
adolescence for ﬁve years, individual psychotherapy in adulthood for two years,
and marital therapy for one year. The other mother participated in individual
psychotherapy in adulthood for three years. Thus, for both mothers in the
resolved group with no previous alternate attachment ﬁgure, they were involved in
individual psychotherapy for a minimum of three years. This is in comparison to
the thirteen insecure mothers with no AAF who, on average, had only three to six
months of individual psychotherapy in adulthood.

In terms of intelligence, the data did not support the prediction that resolved
mothers would demonstrate higher intellectual functioning than insecure mothers.
These comparisons were presented in Table 3. There were no significant
differences on any WAIS-R subtests or in overall IQ between resolved and
insecure mothers, and both groups were within the Average range of intellectual
functioning.

The correlational data displayed in Table 16 and Figure 2 present a
descriptive account of the linkages among alternate attachment figure,
psychotherapy, intelligence, maternal attachment security, and child IQ for

resolved and insecure mothers.

91

 

Table 16
Imgmgclations ﬂpong Child 19, Mother 19, Altemate Attachmept Figpre,
c o a d aterna ttac e t Secu '
AAF AAF Teen Teen Adult Adult Mom
MomIQ Child‘ Teen1 Ther2 Dur’ Ther2 Dur3 Attach

Child 10 .44' .48' .30 .37 .32 .11 .28 -.78”
Mom IQ .33 .15 -.19 -.20 -.25 .03 -.42'
AAF Child .47' -.02 —.19 -.15 -.09 -.53”
AAF Teen .29 .21 -.05 .41' -.61”
Tl‘her .92” .56” .61” -.35
TDur .59” .53” -.29
ATher .29 .10
ADur -.43'

 

1AAF Child and AAF Teen = the presence of a secure alternate attachment ﬁgure in
childhood and adolescence, respectively.

2Teen Ther and Adult Ther = individual psychotherapy in adolescence and adulthood,
respectively.

3Teen Dur and Adult Dur is the length of individual psychotherapy in adolescence and
adulthood, respectively.

(n=26)

'p<.05, ”p<.01

92

:0.v.: . .00.V0 .

 

 

 

 

 

 

 

 

 

 

 

:00u5
“ ‘.[ L 1’ RV.
7 OH EEO _, i
ll ‘. - 1 .:0.-
.3.. .:0.
.- l- :‘7 . t! _
0 11.7.0000... “ - J
«0... , . , ,
. 0 0:03. 00 800...: ,0 .00 F 00 8020: \ @5328... 0......
1‘ - - . .0...-
7 808503.... 0:00:02 7
- l .00. 0...
.00.-_ .
- E 10:00:02 , 0|, - , l.
-. 08.030... .00. ‘ 0% 01! .J
Gm ham-Enh- \ a: 0AQMHOFH L Geog—EU 3 7 Al
F l l , . 1 ﬂ 1

 

 

 

 

 

 

 

 

 

 

 

93

The reorganization of attachment representations in adulthood was related
to the presence of an AAF in childhood and adolescence, IQ, and duration of
individual psychotherapy in adulthood. Of these variables, an AAF in childhood
and adolescence represented the most signiﬁcant correlations to the
reorganizational process, (r=-.53 and -.61, respectively). These negative
correlations indicate that experiences of an AAF (1=yes, 0=no) were strongly
related to low Coherency scores, with low scores on this scale indicating secure
attachment representations. The cluster of psychotherapy variables involving
individual therapy and duration of therapy in adolescence and adulthood were
highly correlated with each other but only the duration of individual therapy in
adulthood was related to the reorganization of attachment representations.
Having an AAF in childhood was signiﬁcantly correlated with an AAF in
adolescence, probably because for many mothers the AAF was a family member
that provided continuity from childhood to adolescence. An AAF in adolescence
was also related to the duration of individual psychotherapy in adulthood. IQ was
not signiﬁcantly correlated to any of the AAF or psychotherapy variables but was
signiﬁcantly related to maternal attachment security. Interestingly, child IQ was
related to the presence of an AAF in maternal childhood, maternal attachment
security, and maternal IQ.

In summary, these data indicated that resolution of issues stemming from
experiences of poor parenting in childhood was strongly related to a secure
attachment relationship with some other adult in childhood or adolescence, and

individual psychotherapy in adulthood for a minimum of two years. For mothers

94

without a previous secure attachment relationship, this resolution process entailed
individual psychotherapy in adulthood for a minimum of three years. In
conjunction with the previous data linking unresolved childhood pain to
dysfunctional second-generation parent-child interaction patterns, these data also
suggested that failure to resolve issues from childhood was related to intellectual
functioning in one’s own children. Supplemental analyses were conducted to

further examine this association.

u e 1 es
Given that adult attachment security was significantly related to maternal

age, education, SES and IQ, it is possible that these variables were responsible for
the relation between maternal attachment security and child IQ. To test this
hypothesis, a post hoc hierarchial multiple regression was conducted. Maternal
IQ and the significant demographic variables were regressed onto child IQ in step
one and maternal attachment security in step two. As Table 17 illustrates,
maternal IQ and the demographic variables were signiﬁcantly related to child IQ.
Collectively, they accounted for approximately half of the variance, although none
of these variables were independently signiﬁcant. When maternal attachment
security was entered into the equation, there was a signiﬁcant increase, accounting
for an additional 15% of the variance. The corresponding beta weights indicate

that of these variables, maternal attachment security was the best single predictor

of child IQ.

95
Table 17

Mpldple Regressjog of Demoggaphic Variables, Maternal IQ and Maternal Attachmegt
Secug'ty on Child IQ

 

 

Bf Adj 3’ _E_ F Change Beta _T_
Child 19
Demo .63 .39 6.69c
Mom IQ .04 .29
Mom Age .05 .25
SES .13 .98
Mom Educ. .20 1.01
Mom Attachment
Reps. .73 .53 908° 11.67c -.47 -3.41c

 

n=
°p<.001

96
Additional analyses were then conducted to ascertain whether quality of

parenting was the mediating link between maternal attachment security and child
IQ. The correlational data displayed in the ﬁrst two columns of Table 18 present
the relations among the parent scales of the PCI, child IQ, and mother IQ. Child
IQ was related to the quality of maternal affect and maternal style of relating in
both play conditions and clean-up. With regard to the quality of parental affect,
higher child IQ was associated with higher maternal positive affect and lower
negative affect. With regard to parental style of relating, higher child IQ was
associated with high maternal positive feedback, encouragement of self-regulatory
and self-reliant behavior, and low maternal parallel play. Additionally, child IQ
was positively related to the degree of parent-child synchrony across all three
conditions.

Parent IQ was associated with the expression of negative affect but not
positive affect, such that higher parent IQ was related to lower scores on parent
negative affect Parent IQ was also associated with style of relating. Higher
parent IQ was related to providing positive feedback, and encouraging child self-
regulatory and self-reliant behavior. Higher parent IQ was also related to lower
scores on compliance to task during the CD play condition and lower scores on
intrusiveness during the PD play condition. Of particular interest is, that
although child IQ was strongly related to parent-child synchrony across all three

conditions, parent IQ was not related to this variable in any condition.

97
Table 18

Intercogelations among Parenting Behavior, Child 19 and Parent IQ.

 

 

Child Mom Partial
IQl IQ2 Corr3

Child Directed Play

Parent Positive Affect .23 .22

Parent Negative Affect -.33' -.24 -.14

Parent Provides Positive Feedback .10 -.08

Parent Facilitates Self-Reg. Beh. .08 -.17

Parent Intrusive .02 -.18

Parent Undercontrolling .05 .02

Parent Demands Self-Reliant Beh. .08 -.05

Parent Compliance to Task -.10 -.30“

Parent Parallel Play --- «-
Parent-Child Synchrony .59” .19 .57‘
Parent Directed Play

Parent Positive Affect .29‘ .19 .18

Parent Negative Affect -.21 -.33‘

Parent Provides Positive Feedback .12 .08

Parent Facilitates Self-Reg. Beh. .12 .27

Parent Intrusive -.27 -.44‘

Parent Undercontrolling .05 -.26

Parent Demands Self-Reliant Beh. -.16 -.14

Parent Compliance to Task -.01 .04

Parent Parallel Play -.37' -.35' -.10
Parent-Child Synchrony .45” .20 .42”
Clean-Up

Parent Positive Affect .33‘ .12 .29

Parent Negative Affect -.16 -.37"

Parent Provides Positive Feedback .41” .43” .20

Parent Facilitates Self-Reg. Beh. .49” .43” .30"

Parent Intrusive -.05 -.13

Parent Undercontrolling -.05 -.02

Parent Demands Self-Reliant Beh. .29‘ .30' .21

Parent Compliance to Task .02 .18
Parent-Child Synchrony .37' .13 .33”

 

110 derived from Stanford-Binet abbreviated test battery.

2IQ derived from Silverstein (1982).

3Partial corr between Child IQ and parenting behavior, controlling for Mom IQ
N=46

'p<.05, ”p<.01, °p<.001

98

These results indicated that while child IQ was related to particular
expressions of maternal affect and parenting behaviors, the most consistent
association was between child IQ and parent-child synchrony. Further, parental
ability to engage in synchronous interactions was determined more by
representational models of attachment than intellectual functioning.

These ﬁndings were underscored by the partial correlations between child IQ
and parenting, controlling for maternal IQ, as presented in column three of Table
18. The partial correlations indicated that for the parenting scales signiﬁcantly
related to child IQ, the only speciﬁc parenting behavior that remained
signiﬁcantly related to child IQ was facilitating self-regulatory behavior, once the
effects of maternal IQ were taken out of the equation. Yet, the degree of parent-
child synchrony remained significantly related to child IQ across all three
conditions.

In summary, results from the multiple regression and correlational analyses
indicated that the manner in which mothers mentally organized and reﬂected
upon their own childhood histories had a direct impact on parent-child
interchanges involving parental expectations for child self-regulatory behavior and
child compliance, and on the degree of synchrony in parent-child interactions. In
addition, maternal attachment security imparted a unique inﬂuence on child
intellectual functioning, above and beyond that which could be explained by the

more conventional variables of parent IQ, age, education and SES.

99
Hypdtheses 5

There was partial support for the prediction that mixed parent-child dyads
would fall mid-way between secure and insecure dyads with respect to parent-
child interaction patterns. An exploratory AN OVA and Bartlett’s Box Test
revealed that mixed mothers were more similar to the insecure mothers than to
the secure mothers on the parent-child interaction task, but also had signiﬁcantly
more heterogeneity of variance than either group. This precluded further
analyses of these individuals as a separate group. Therefore, for additional
exploratory analyses, seven of the individuals in this group who had received
Loving scores of 3 and 4 and Coherency scores of 4 were included into the
insecure group on the basis of their Coherency scores. ANOVA analyses
comparing the newly formed combined insecure group (11? 22) with the combined
secure group (2 =21) were then conducted. Table 19 presents these comparisons
and indicates a slight shift in the differences between dyads with secure and
insecure maternal attachment security.

A comparison of Tables 8 and 19 illustrates that combining mothers with
mixed and insecure attachment security enhanced the "positive" differences and
diminished the "negative" differences between the original combined secure and
insecure attachment groups. For instance, in the CD condition, the previous
ﬁndings regarding likability of the child and parent-child synchrony in the
combined secure dyads were replicated. The differences were that, in addition,
combined secure mothers encouraged more self-reliant behavior from their

children and were rated as more likeable than combined insecure mothers. Also,

100

the previous pattern of increased negative affect and distress in children of
insecure mothers was reduced below a level of signiﬁcance in the combined
insecure group. A similar pattern emerged in the PD condition. Parent-child
synchrony remained higher in the combined secure dyads compared to the
combined insecure dyads. In addition, combined secure mothers demonstrated
more positive affect, provided more positive feedback, and were rated as more
likeable than combined insecure mothers. Also, although parent negative affect
and child disobedience remained signiﬁcant in the combined insecure group, it
was less so and parent intrusiveness and parallel play were reduced below a level
of signiﬁcance. In the CU condition, parent positive feedback and
encouragement of self-reliant behavior remained signiﬁcantly higher in the
combined secure group than the combined insecure group. Differences included
an increased level of signiﬁcance in child compliance, activity and parent-child
synchrony in the combined secure dyads compared to the combined insecure
dyads. Further, combined secure mothers were rated as more likeable then
combined insecure mothers. There was also a non-significant trend, suggesting
that combined insecure parent-child dyads required a longer time to complete
clean-up than combined secure dyads.

With regard to child outcome, the results supported the prediction that
children of mixed mothers would fall mid-way between children of secure/resolved
mothers and children of insecure mothers with respect to intellectual functioning
and attachment security. Table 20 presents these comparisons. In terms of

intelligence, children of mixed mothers were significantly higher than children of

101

insecure mothers on Verbal Reasoning, Memory, and overall IQ. Further, there
was a non-signiﬁcant trend, suggesting that children of mixed mothers scored
lower than children of resolved and children of secure mothers in these same
domains. In terms of child attachment security, a similar pattern emerged.
Children of mixed mothers were signiﬁcantly more insecure than children of
resolved mothers on the Spilled Juice story. Further, although the means were
not statistically signiﬁcant, children of mixed mothers were rated as more insecure
than children of secure mothers and more secure than children of insecure
mothers on this scale. Similarly, on the Hurt Knee story, children of mixed
mothers were rated as signiﬁcantly more secure than children of insecure
mothers. Once again, although the means were not signiﬁcantly different, they
were higher than the means for children of resolved and secure mothers. Finally,
with respect to their overall security score, children of mixed mothers were mid-
way between children of resolved/secure mothers and children of insecure

mothers, although these differences only approached signiﬁcance.

Summag of Results

Overall, the data supported the predicted model. With regard to replication
processes, women who had experienced painful and unloving parenting in
childhood but had resolved these early experiences were similar to women who
had experienced loving and secure childhood relationships with respect to
parenting in adulthood. While self-report data suggested some residual effects of

childhood experiences on adult intrapersonal functioning and interpersonal

102

:00.v.:.. .:0.v.:. .00.v.:. .0:.V0.

5:80:

:0 :0 5:": :33 A2000 55:: m 0 :0 0000:: 0:0 :05? 00:0: :0=0:00: 0:0 3:009:00 0:30.: 5: 50000: 0:000 .50.: n 0 :0 0000:: 0:002

ANNHE 05000:: 00:50:00 0:0 Gmué 0::00m 00:30:00

 

 

 

 

 

 

 

 

DU :: 00000 :0: 0:00w 80 8. 8.: 8. 8.: --- 8. 8.: 8. 8.:

8.: bn. 8.~ mm. 80 .8.” 8. 80 8. 00.0: 8.: 8. 8.m 0b. 00.0-N
00:... 8. 8.: 0:. 8.: mum 8. cm: 0.0. mm: .:00 m:. ::.: 0:. NN.:
8. 8. 8.: No. 8.: cm. 8. 8.: No. 8.: 00. No. 8.: no. 8.:
ma. 0:. 8.: 8. 8.: .80: mn- cm: w:. 8.: mm. m:. 8.: 8. 8.:
ﬁnd em. «:0: mm. wmd .2: 0N. :00” 8. 0m.»- 0:.~ 0m. 0N6 0N. Swm
90.0.0 on. 8.N mm. w~.~ 08.»: Va. mad mm. mn.~ 3.N 0N. 0N.N 0m. :00
8.: 0:. 8.: 8. 8.: .86. n:. 8.: N0. 8.: SN 00. 8.: 8. 8.:
0:0 9.. $0 8. 80 .86 9.. Run on. 80: : m: 0m. cod mm. GEN

0:03: 030.: 0:00”:
W. le :002 Dlm 82 M NW :002 dum- :002 N ﬁlm :002 le :002
000:: 0:50 08 0:50 000:: 0:50 00m 0:50 000:: .0500 00w 0:50
0:3 yam-:0 >50: QEmED .:.Zm:~:<.: >50: OFF-Oman: Q1:=:U

00:0 0:90.:
8:020:50
._0~:.::0m
.:00:00:D
02035:
.000-000
.0000: .00.:
80:2 .87:
000:2 .00.:
a

 

 

 

5085034.: 35202 0.50005. 00:30:: 00 0:0 0::00ﬂ0050800 :00: 05020.: :o::00:0::: 0=no-50:00:

a: 030:.

80.x... .:0.V0. .00.V0.. .0:.v0.
::.:umw
::0 :0 00:": 5:3 6:000 :50: n 0 :0 00000 0:00:

ANNHS 05000:: 0030800 0:0 A:~u:v 0.500m 00:50:00

 

103

 

 

 

 

 

 

 

 

. 0.0 :0. 00.0 00. 0.0.0 .000: 00. 00.0 00. 00.0 .000: 3.. 00.0 00. 00.0 0800 0-0
0:0 00. 00.0 00. 0:0 0:. 3. 0:0 00. 00.0 00. :0. :00 00. 00.0 303..
00.0 00.: 00.0 00. 0.0 00.0 00. 00.0 00. 00.0 .000 R. 00.0 00. 00.0 20: 2:00
.00... 0:.: 00.0 00. 00.0 00.0 00. :00 00. 00.0 .000 00. 00.0 00. 0:0 80: .0200
00. 00.: 00.0 00. 00.0 00.: :0. 00.0 00. 00.0 00.: :0: 00.0 00. 00.0 000000:
g
00. 00. :0: 00. :0: 00. 00. 00.: 00... 00.: 0:.: 00. :0: 00. 00.: 0808000
00. ::. 00.: :0. 00.: .00... 00. 00.: 00. :0: 0:.. 00. :0: 00. :0: 0385
.000 00.: 00.0 00. 00.:. 00.: 00. :0: 00. 00.: 00.: 0:. ::.: 0:. 0:.: 8:00:50
:0. 0:.. 00.0 00. :00 00. 00. 00.0 0... 00.0 00.: 00. 00.0 c.. 00.0 00508.:
00. 0:. 0:.: 00. 0:.: 00.: ::. 00.: :0. 00.: .:00 00. 0.0.: 00. 00.: 002.00
00. 0:. 00.: 0.0. 00.: 00.: 00. 00.: 00. 00.: .000 00. 00.: .00. :0: .800. .002
:0. 00. 00.0 00. 0.0.0 0:. 0:. 00.0 00. ::.0 00. 00. 00.0 0:. 00.0 080.0. .80
00:0
000:: 000:: 000::
m. a :002 mm. :00: M mw. :002 NW. :002 .0: NW: :002 a :002
82.: .0080 80 .0800 88: .0080 80 .0050 88: .0080 80 .0600

 

.:D Z<m:..:0 >4: QEDm::Q Hzmga: >50: QEWKR: QA:=0

30:30.33 ohaoomﬁuva 28 25:32:? ﬁguNA .23 .23» 2:2. 9 a no v8.3 8:03 89)
ASE“:— mutoa 89A 82th 0A .883 «oz Eng—3m 9:89:00 Aoumméheqﬁm 80:2 2:32“
mAaovA-§EBoZJ=oA.3m "VAZP

gang ans 352 5 us 2835 S_us 28% Ens 328;

104

 

 

 

 

 

 

 

A ”1A.“? no. 3d 8A mm.~ we. 8N cw. 8N mu. wA.N EoﬁﬁmS<
hm. ac. hm. $.N an. MEN 5AA 3N 2: had now—33A
9.. co. MA A omﬁ mm. EN 3. SN 2.. ANN cosy—«mom
mo. 5A. NAA mm.~ a. mnﬁ S A San NcA 3.N 33.52
AV EwJA 8. And mm. 3N we. c~.m co. A om.N no. 8.~ 35A tam
33¢ VAA S. Rd m A . A 38 an. $.N 3. 8:0. Aw. ch "85—. Ago—Ea
2 Ad
AAszA A8. NadA mnd 3.2: cm. AA maﬁa 36 86A A 86 Qua A CA
AA «4m;— AS. Amé ANNA 8.2: gum A 5.3 cm. A AA caNA A 8.3 modNA .0080:
ac. va mud 3.2: 8.3 $63 EBA mm. A AA nNdA 86A A 45:0
AAmJA A... 8.9 E .w 8.2: owA A 3.8 Awd cNA AA ood 36: Subunxx
A A 2
Ad Ad A8. 8A. Nmﬁ A 8.2 A and A 8.8 om.mA 3.3 A mad wA .NNA «Ana._0>
_lemAumu 33 033A ”A ﬁlm :32 ﬁlm: 532 dim. 532 a :32
Oman—2 mmammzA mmDUmAm QmA>AOmmA~A

 

 

 

cm 033.

105

relationships, these data were speculative. In addition, observational data
indicated no residual effects from the parenting experienced to the parenting
provided for the resolved mothers. Resolved and secure mothers expressed more
warmth, affection and enjoyment when playing with their children and they
encouraged more self-reliant behavior from their children than insecure mothers.
In turn, children of these mothers were more compliant during clean-up and were
rated as more likeable than children of insecure mothers. Finally, these dyads
engaged in more synchronous interactions than insecure dyads during both play
and clean-up conditions. Conversely, insecure mothers were more hostile,
negative and sarcastic in their affect. They were more intrusive, manipulative and
overcontrolling in their play, and were more likely to engage in emotionally
disconnected, parallel play. The children of these mothers expressed more anger,
dislike and/or hostility toward their mothers, demonstrated more emotional
distress, and were more disobedient than children of combined secure mothers.
A variation on this theme was the quality of parenting demonstrated by mothers
with mixed attachment representations. These mothers seemed to reenact their
experiences of adequate parenting, as evidenced by the absence of the positive
parenting qualities characteristic of the secure mothers and the negative parenting
qualities associated with the insecure mothers.

Also in terms of replication patterns, child attachment security and
intellectual functioning were associated with maternal attachment security.
Children of resolved and secure mothers were more likely to demonstrate secure

representational models of attachment compared to children of insecure mothers

106

and both groups scored higher across several intellectual skill areas compared to
children of insecure mothers. It is interesting that children of resolved mothers
demonstrated higher intellectual functioning than children of secure mothers,
despite the fact that resolved mothers did not demonstrate higher intellectual
functioning than secure mothers. Since child IQ was strongly related to parent-
child synchrony, this suggests that resolved mothers, in an effort to provide
experiences for their children that they themselves did not receive, engaged in
even more positive interchanges with their children than mothers who had loving
relationships with their parents. This finding also poses provocative implications
regarding the role of psychological and social factors in intellectual development
With regard to the reorganization process, the experience of a secure
alternate attachment figure at some point in the early developmental history was
strongly related to resolving the pain stemming from unloving parent-child
relationships. Individual psychotherapy in adulthood was also related to this
process, with two years representing the average length of this therapy, and
intelligence was unrelated to psychotherapy experiences. Although maternal IQ
was related to maternal attachment security, resolved mothers did not
demonstrate higher intellectual functioning than insecure mothers. These findings
suggest that above average intelligence is not requisite for benefiting from
psychotherapy and/or resolving issues from one’s past. Finally, given that the
majority of resolved mothers who participated in this study had female children, it

suggests that having a daughter may facilitate the process of resolution.

107

In any event, it is clear that mothers who have not resolved the past, go on
to recreate dysfunctional parent-chﬂd relationships as they embark on parenting.
This relation between unresolved conﬂicts from childhood, current parenting, and
child attachment security is consistent with established theoretical models and
other research. The connection between maternal resolution of the past and
subsequent child intellectual functioning is an intriguing finding that represents

new territory.

CHAPTER FOUR
Discussion

This study provided meaningful data about the replication of dysfunctional
parent-child relationships, experiences underlying discontinuity in developmental
trajectories within and across generations, and the impact of maternal functioning
on child outcome. It also identified several methodological issues for further
investigation in this field of inquiry. An interpretive discussion of the major
issues and implications posed in this study is organized around the following
topics: methodological considerations; replication processes; reorganizational
processes and residual effects; and intelligence. Finally, limitations of the study

and directions for future research conclude the discussion.

Methodological Considerations

The findings of this study highlight four major methodological issues
related to the study of attachment security throughout development. The first
issue involves the validity of the AAQ as a viable research alternative to the AAI
for classifying adults broadly aseither secure or insecure with respect to their
representational models of attachment. The inter-rater reliabilities reported in
this study and in Adams (1992) suggest that the modified scoring system
demonstrates adequate reliability for screening purposes. Although the modified
version still requires intensive training, it is much less than that involved with the
AAI and original scoring system. The current study indicates that this training

can be successfully accomplished by following the manual and within a reasonable

108

109

period of time once coders are familiar with generic issues of attachment theory
and defensive processes in psychological functioning.

The discriminant validity of the AAQ is not as straightforward, as there
was some relation between intelligence and attachment security classifications.
This finding is in contrast to that reported by Bakermans-Kranenburg & van
Ijzendoorn (1993) who found that adult attachment security, as assessed with the
AAI, was independent of intelligence. There are a number of possibilities
accounting for the discrepancy in these findings. Perhaps the written descriptions
of one’s history offered by the AAQ are more dependent upon verbal ability and
overall intelligence than the oral presentations provided by the AAI.
Alternatively, it may be that the modified scoring system is not as discrete in
identifying attachment status as the original scoring system. It is also possible
that without formal training in the original scoring system, nuances and subtleties
are overshadowed by the more global constructs of verbal expressiveness and
general reasoning ability. Also, Bakermans-Kranenburg & van Ijzendoorn utilized
different measures of intelligence and sampled subjects from a different culture
than the current study so it is difficult to make comparisons. Finally, the sample
size of this study was relatively small so the relation between IQ and attachment
security may reﬂect a sampling bias. Future studies are needed to clarify these
issues. In particular, it would be helpful to use the AAQ with the original scoring
system and the AAI with the modified scoring system with larger and more

culturally diverse samples.

110
There are additional limitations of the AAQ and questions that need to be

addressed. For example, it is not clear whether the AAQ provides enough
detailed information to classify individuals by type of insecure attachment (i.e.,
dismissing, preoccupied). This may represent a significant limitation, depending
on the nature of the research question. In addition, there are potentially
numerous variations to the method of administration and it is not clear what the
impact of these variations might be on the quality of responses. For example, in
the current study, subjects were mailed the AAQ to complete in their homes
whereas Adams (1992) employed a group administration on a university campus.
On the one hand, the group setting afforded more control over the manner in
which the subjects progressed through the questionnaire but it also may have
created an impersonal, public atmosphere that inhibited subjects’ responses.
Having subjects complete the questionnaire in their own home is more efficient
from a practical standpoint and it may increase their level of ease and comfort in
responding to items. However, there is no way to monitor whether subjects
answer the questions in sequence or within one sitting. This may be important as
the AAI was designed to probe subjects’ attachment representations with
unexpected and progressively difficult questions. Yet, the results from the current
study suggest that the responses subjects provided reﬂected their representational
models of attachment, despite not being able to control for these variables.
Finally, subjects in the current study and in Adams (1992) were predominately
middle to upper-middle class, Caucasian women. It is unknown how much these

results can be generalized to other samples varying on these demographic

111
variables. Additional research is needed with the AAQ to address some of these

questions and provide further validation for its continued use in research.

Despite these limitations and cautions, classifications of adult attachment
security derived from the AAQ were meaningfully related to parenting behavior,
child behavior, child attachment security, and child intellectual functioning.
Furthermore, these relationships were uniquely related to maternal attachment
security independent of maternal age, education, SES or IQ. As such, this study
provides preliminary data about the potential contribution of the AAQ for future
research efforts.

The second methodological issue relates to the use of self-report measures
in studies involving adult attachment security. The fact that mothers with
insecure attachment representations did not differ significantly from mothers with
secure attachment representations on a self-report measure of distressed parent-
child relationships but demonstrated more distressed parent-child relationships on
the observational measure indicates that individuals with insecure attachment
security are not accurate reporters of problems. This is consistent with other
research that has reported similar discrepancies for individuals with insecure
attachment, particularly of the dismissing type (Cohn et a1., 1991; Kobak &
Sceery, 1988; Spieker & Booth, 1988). This is likely due to the pattern of denial
and failure to attend to painful and/or negative affect that are hallmark features
of the avoidant/dismissing attachment style. This element is inherent in the
scoring system devised by Main & Goldwyn (1989) who note that dismissing

adults characteristically have idealized and defensive presentations of attachment

112
relationships. The current study provides additional evidence that individuals with

insecure attachment representations lack an awareness of the difﬁculty present in
their lives and as such, do not identify problems on self-report measures of
distress. An alternative explanation for this finding is that the AAQ misclassified
individuals with respect to their attachment security but given the theoretically
consistent relations among maternal attachment security, parenting behaviors, and
child functioning, this is not a likely possibility. Aside from the implications this
finding has for research, it is particularly relevant from a clinical point of view.
Resolved mothers, who demonstrated very positive parenting behaviors, tended to
report concern and stress about their parenting abilities whereas insecure mothers
reported as much confidence in their parenting abilities as secure mothers.
Therefore, the parents most in need of intervention were the least likely to
identify themselves as such. This provides empirical support for the theoretical
assumption that insecure attachment representations are highly organized internal
models that operate outside of awareness and thus perpetuate dysfunctional
parent-child relationships. This speaks to the difficulties inherent in intervention
efforts. Helping these individuals recognize the usefulness of therapy would be a
major task and the course of treatment would necessarily be an in-depth process
as the presenting model would need to be slowly challenged and dissembled while
gradually reconstructing a more adaptive one.

In light of the discrepancies between self-report and observational
measures of distressed parent-child relationships, the third methodological issue

raised in this study is the need for behavioral observations in identifying

113
disturbances in parent-chﬂd dyads. With respect to research, the modiﬁed version

(Whipple et a1., 1993) of the Belsky Parent-Child Interaction Coding System
(Belsky et a1., 1990) is effective and practical for categorizing adaptive and
maladaptive parent-child interaction patterns. However, there was a curious
finding regarding the Self-Reliant scale on this measure as it relates to parenting
behavior. In both scoring manuals, the Self-Reliant and Self-Regulatory scales
reﬂect parental behavior that is directed toward promoting child autonomous
behavior. The difference is that the Self-Reliant scale lacks the element of
facilitation or "scaffolding" to support the child’s efforts. Given that parents who
scored high on the Self-Reliant scale also scored high on parent-child synchrony,
it appears as though this difference was not incorporated into the coding in the
current study. Consequently, the Self-Reliant and Self-Regulatory scales appear
to have tapped similar parental behaviors. This should be addressed in future
studies by giving additional emphasis to the difference between these two scales in
the coding manual and in training. Finally, since child behaviors were
predominant during the child directed play activity and parent behaviors were
predominant during the parent directed play and clean-up activities, the study
provides empirical justification for this task differentiation in assessments of
parent-child interactions. This finding also suggests that parents provide a more
inﬂuential force in parent-child interactions for goal-directed activities. This is
consistent with the findings discussed by Crowell & Feldman (1988), who suggest
that issues involving child anger/aggression, disobedience, and compliance result

from parenting behaviors rather than the obverse. This raises implications

114
regarding the initial impetus in coercive interactional cycles in parent-child
relationships.

Finally, the fourth methodological issue relates to assessing child
attachment security on a representational level in pre-school age children. In
general, the present study suggests that the ASCT is an internally consistent
measure of representational models of attachment and that child attachment
security is meaningfully related to maternal attachment security and child
intellectual functioning. However, there were problems identified with this
measure. While the overall attachment score was significantly related to maternal
attachment security, only two of the five stories independently demonstrated this
relation. These stories involved themes of discipline and injury as elicitors of
attachment dynamics. This finding was surprising given that Bretherton et al.,
(1990) state that the stories involving separation and reunion with parents should
carry more weight than the other stories. There are several possibilities for this
finding. Without specific training and/or a detailed scoring manual, it was very
difficult to know to what extent my interpretation and administration of this
measure was consistent with the intent of the authors. Bretherton (1994, personal
communication) has commented on the preliminary nature of the instrument in
its present form and is currently working on a revised and more detailed scoring
manual. This will undoubtedly facilitate future research efforts. In addition,
since 94% of the children in this study were living in two-parent families, it is
quite conceivable that paternal attachment security exerts a significant inﬂuence

on child attachment security in ways that may either be cumulative or

115

contradictory to maternal attachment security. Without this information, it is
difﬁcult to know to what extent the absence of a more significant effect between
child and maternal attachment security reﬂects measurement error and/or missing
data. These questions need to be pursued in additional validation studies of the
ASCT.

Another interesting facet of the ASCT involved maternal presence. As
stated earlier, the ASCT was typically administered immediately following
maternal separation from the child in the playroom and the majority of the
children in this study did not have difficulty separating from their mothers for this
assessment. However some children could not tolerate the separation and
required that their mothers remain present during the entire assessment.
Although there were no signiﬁcant differences in the children as a function of
maternal attachment security, the descriptive data revealed that only children of
insecure mothers required maternal presence during the ASCT. Thus, the only
children who had problems separating belonged to mothers who had difficulties in
their own attachment security. This finding may have implications for children

with Separation Anxiety Disorder.

Replicagon Processes

The major thrust of this study provides empirical evidence for the role of
adult attachment representations in the transmission of dysfunctional parent-child
relationships across generations. Specifically, it suggests that the manner in which

childhood histories are mentally organized and integrated in adulthood impart a

116

profound inﬂuence on parent-child behavior patterns in the next generation. This
link is manifested in the quality of parental affect and style of relating. In
combination, these two dimensions of parental behavior create two distinct
behavioral patterns. Mothers who had experienced loving and secure
relationships with their parents in childhood, or who had come to terms with
unloving and painful relationships, expressed warmth, affection and enjoyment
when playing with their children. Further, their style of relating was to provide
more positive feedback, and to facilitate more self-regulatory and self-reliant
behavior from their children. In addition, mothers and children in these secure
dyads engaged in more synchronous interactions during play and clean-up
activities. Conversely, mothers who had experienced unloving and painful
relationships with their parents during childhood and had not resolved these
issues demonstrated hostile, negative and/or sarcastic affect when playing with
their children. Additionally, their style of relating was intrusive, overcontrolling,
and/or emotionally disconnected. In turn, their children expressed anger, dislike
and hostility toward them, emotional distress, and disobedience during play and
clean-up tasks. The fact that these patterns were evident in a community-based
sample, during a videotaped, low-stress condition of free-play at a university
campus is of enormous significance. One can imagine how these patterns may
become more pronounced in conditions of high stress and/or low visibility that is
more characteristic of home settings.

While these data provide compelling evidence for the role of adult

attachment representations in parent-child interactions, there were complex

117
relations among this variable and speciﬁc demographic variables. Adult

attachment security was highly correlated with maternal age, education, and
socioeconomic status and secure mothers were significantly higher on these
variables than insecure mothers. These results are somewhat similar to those
reported by Bakermans-Kranenburg & van Ijzendoorn (1993) who found that
secure mothers were signiﬁcantly older than preoccupied and dismissing mothers,
and that unresolved mothers had the lowest levels of education. One explanation
for this pattern of results is that individuals with secure attachment in adulthood
demonstrate higher levels of competency than insecure adults in the
developmentally equivalent tasks of intimate relationships, education and career
pursuits, just as attachment security is strongly associated with competency in
childhood (Erickson, Sroufe & Egeland, 1985; Grossman, Loher, Grossman,
Scheuerer-Englisch, Schildbach, Spangler, Wensauer & Zimmerman, 1993; Sroufe,
1983) and adolescence (Kobak & Sceery, 1988). Thus, secure adults may pursue
higher levels of education that would lead to more advanced occupational
achievements and postpone the life cycle stages of marriage and child-bearing.
With more emotional, interpersonal and financial resources, their lives are less
stressful and they demonstrate more positive emotions and behaviors when
interacting with their children than mothers without these benefits. In addition to
this general pattern, the manner in which mothers have internally organized their
childhood experiences of being parented imparts a particular inﬂuence on a triad
of parent-child behaviors involving parental expectations for child self-regulatory

behavior, child compliance, and the degree of parent-child synchrony.

118

The question remains as to how the quality and nature of parent-child
interactions are involved in the replication of relationship patterns across
generations. In the present study, the strongest and most consistent difference
between mothers of secure and insecure attachment representations was the
degree of synchrony in the mother-chﬂd dyads. That is, despite the fact that
secure mothers elicited more self-regulatory and self-reliant behavior on the part
of their children than insecure mothers, and children of secure mothers were
more active, the secure mother-child dyads consistently engaged in more
synchronous interactions across play and clean-up conditions than insecure dyads.
The proposal advanced in this study is that this experience of dyadic synchrony is
the core element underlying the continuity of relationship patterns across
generations.

The notion of parent-child synchrony is implicated in numerous other
studies and theoretical models of development (Emde, 1983; Fonagy, 1993;
Fonagy, Steele, Steele, Moran & Higgitt, 1991; Murray, 1991; Osofsky, 1993;
Stern, 1985), with varying interpretations as to the specific nature of its inﬂuence.
Murray (1991) comments on the caregiver’s capacity for "entering the dyadic
circle of engagement and meeting the infant’s communication" (p. 222), and the
impact of this process on the infant’s capacity for internal affect regulation.
Emde (1983) posits that as the child’s affective core becomes organized, the
process of self awareness and development evolve through experiences of
"behavioral synchrony” and "communicative connectedness" which are rooted in

the emotional availability and responsivity of the primary caregiver. Fonagy

119
(1993) and Fonagy, et al., (1991) argue that the ”reﬂective-self function," a crucial

self structure underlying representational models, is embedded in the infant-
caregiver relationship and evolves from experiences of attunement in infancy to
”empathic sensing and responding" in early childhood, and that this structure is
meaningfully related to attachment security across generations. Osofsky (1993)
discusses the implications of ”dyadic regulation and dysregulation" for the child’s
ego development, whereas Stern (1985) has written extensively on the quality of
affective exchange or "attunement" between parent and child and its impact on
the child’s development of self.

These theoretical models and previous work on parent-infant interactions
indicate that optimal development entails emotional interchanges between parent
and child whereby the parent shares in the process of experiencing, expressing,
and responding to affective cues within herself and in relation to the child that
creates a climate of shared emotional experience between them. Data from the
current study suggest that parental ability to engage in this process is meaningfully
related to representational models of attachment. Parents with insecure
attachment representations, by virtue of their distorted, repressed and defensive
strategies, project these qualities into the behavioral transactions with their
children. This subsequently restricts and binds their ability to acknowledge and
respond to particular affective expressions and behavior on the part of their
children. This was especially evident in the child-directed play condition where
mothers were required to follow their child’s lead. The scores on parent-child

synchrony indicate that insecure mothers had the most difﬁculty in this condition.

120

Therefore, to the extent that mothers must employ a defensive organization of
thought with respect to their own childhood histories, they create disjointed and
dysynchronous patterns of interaction with their children. In turn, the children
are prohibited from integrating certain affective experiences and behaviors as
their self organization evolves, and they learn to employ similar defensive
strategies.

The interaction patterns that emerged in this study provide support for this
model. However, it should be noted that while these patterns were distinct and
statistically signiﬁcant, they were subtle from an observational standpoint. An
untrained observer would most likely be unable to identify behavioral differences
between secure and insecure parent-child dyads. Thus, it may not typically be the
case that parents with insecure attachment representations commit blatant and
atrocious actions that violate, denigrate or invalidate their childrens’ sense of
being and experiencing. Rather they emit low-level, chronic and pervasive
behaviors that convey negative messages to and about the child, intrude upon
and/or dismiss crucial affective experiences, and thus distort the interactive
structure. It is suggested that this deficit in interactive regulation becomes
internalized and incorporated into the child’s sense of self and other and the
interplay between these two core psychological structures is carried forward in
subsequent relationships.

One of the most discouraging aspects of this type of repetition process is
the powerlessness of intent. Many of the insecure mothers in this study were

cognitively aware of the problems they experienced with their own parents but

121

were split off from the affect associated with these memories. They provided
rational comments about how they vowed not to repeat the mistakes in parenting
their children that they experienced. Often with great pride, they discussed how
they had broken out of the cycles from their families or origin. While it remains
a distinct possibility that there was significant improvement from the parenting
they received to the parenting they provided, they were still engaged in
destructive interaction patterns with their children, despite their earnest intent not
to recreate the past. This provides further evidence that accessing and integrating
childhood pain is a deterrent against repetition in parenting, while repression,
isolation, or absorption in painful affect are linked with reiteration of these
patterns with one’s own children. Thus, cognitive awareness and intent, in the
absence of integrated affect, are insufficient coping strategies for breaking free
from the patterns of the past. This has profound implications for intervention

efforts.

Reorganizational Processes and Residual Effects

One of the most inspiring aspects of this study is the degree to which
individuals can redirect the course of their lives. Many of the resolved mothers
had experienced severe physical, emotional and/or sexual abuse, physical and
emotional abandonment, parental alcoholism and depression. Despite these
adverse experiences, there were no residual effects of the parenting they
experienced on the quality of parenting they provided. In fact, they demonstrated

positive parenting behaviors equal to and above those of parents who came from

122

loving and secure backgrounds. This discontinuous trajectory from childhood to
adulthood is a fascinating issue that is addressed by others authors (Egeland,
Jacobvitz & Sroufe, 1988; Quinton & Rutter, 1988) and speaks to issues of
change, self-righting tendencies, recovery, and resiliency in human development.

In the current study, the most signiﬁcant experience related to the
reorganizational process in adulthood was the presence of a secure alternate
attachment ﬁgure in childhood or late adolescence. The overwhelming majority
of resolved mothers had experienced this kind of relationship compared to almost
none of the insecure mothers. In terms of impacting the reorganization of
attachment representations, it did not seem to matter whether this relationship
occurred in childhood or adolescence, suggesting that the consolidation of
representational models remains malleable at least until early adulthood.
However, since more than half of the resolved mothers experienced an alternate
attachment relationship in childhood, it is not known whether the intensity and/or
duration of this relationship must increase dramatically as development progresses
in order to achieve a similar inﬂuence. What is clear is that the experience of
being genuinely loved and nurtured has an enormous impact on one’s sense of
self and contributes powerfully to the capacity for subsequent caregiving. It
suggests that from the experience of being nurtured evolves the capacity for
nurturing.

Individual psychotherapy also appears to play a signiﬁcant role in the
process of reorganization. It may be that this experience provides a context for

integrating the affect associated with the past, as well as creating a sense of being

 

 

123
respected, cared for, and tended to that serves to challenge the representations of

self and other that developed in response to negative parent-child relationships.
It is interesting that engaging in psychotherapy in adulthood was not related to
one’s representational model of attachment, but that the duration of the
therapeutic relationship was. Resolved mothers were involved in individual
psychotherapy for an average of two years, compared to insecure mothers who
participated in therapy for an average of three to six months. As noted in other
studies, individuals with insecure attachment representations are typically involved
in interpersonal relationships marked by intense emotional conﬂict, distrust and
dissatisfaction. This suggests that their difﬁculty to engage successfully with
others may extend to therapeutic relationships, precluding the opportunity for
them to encounter a "corrective emotional experience". Resolved mothers, on the
other hand, because of their alternate attachment relationships have a positive
relationship experience to draw from as they engage in therapeutic relationships,
thus enabling them to sustain the level of intimacy requisite for altering their
representational models. While this sequence may represent the modal process, it
is also clear that one can achieve a similar outcome without the experience of a
previous alternate attachment figure. However, in those cases, the duration of
the therapeutic relationship appears to increase significantly.

Another intriguing finding of this study involved child gender. It was quite
unusual that almost all of the resolved mothers had a female child. This raises
the possibility of the child’s role in the reorganization of maternal attachment

representations, perhaps through the process of projective identification.

124

Goldstein (1991) provides an overview of the concept of projective identiﬁcation
as it was initially conceived and how the meaning and application of this construct
have evolved over time. The basic model of projective identiﬁcation involves
three stages. In the first stage, a rejected aspect of oneself is projected onto
another person. The second stage involves interpersonal interactions as the
projector attempts to make the recipient act, think, and feel in accordance with
the projection. Throughout the course of these interactions, the projector is
provided with opportunities to psychologically process the rejected material. This
leads to the third stage whereby the projector reinternalizes the self
representation. Goldstein distinguishes between a severely pathological form of
projective identiﬁcation and a more benign, commonplace one. In the first case,
there is blurring of the boundaries between the projector and recipient as
internalized drives are projected. In the second case, the individual projects
object representations that do not involve this kind of psychological merging. It is
the latter process that Goldstein argues is ubiquitous in interpersonal
relationships. Also, Goldstein notes that in stage three, the degree of
psychological processing and reintemalization can range from primitive to
sophisticated and adaptive forms, depending on the maturational level of the
projector. In this way, the process of projective identification offers the potential
for attaining new ways of handling threatening feelings and fantasies and this
mastery enables the individual to integrate the previously rejected aspects of the
self. Although projective identification has typically been related to the process

of psychoanalytic psychotherapy, Goldstein affirms that it is being expanded to

125
incorporate the dynamics underlying other interpersonal relationships, including
parent-chﬂd relationships.

In the context of the current study, the process of projective identiﬁcation
may offer further insight about the reorganization of attachment representations.
Although all children can be potent elicitors of unresolved issues for adults,
female children may play a particular role for mothers attempting to come to
terms with their past. In other words, perhaps the potential for projective
identiﬁcation is enhanced in mother-daughter relationships because of the gender
replication. In this manner, mothers may project their attachment issues (i.e.,
emotional and physical dependency) onto their infant daughters. Since these
experiences are syntonic to the needs of the infant, it does not necessarily involve
overriding the needs of their children that would otherwise create distorted
interactions. In the course of confronting the issues within this framework,
mothers may be able to process and reinternalize them in a more adaptive form.
However, as Goldstein pointed out, this outcome is dependent upon the level of
psychological maturation in the projector. Given that resolved mothers had
experienced an average of two years of therapy prior to having had their children,
they may have been prepared to engage in this process in a way that was not
possible for insecure mothers.

Although the presence of an alternate attachment figure, individual
psychotherapy, and having a daughter all seem to be tied into the reorganization
of attachment representations, it is still not clear to what extent these experiences

represent independent pathways or work in tandem to help individuals resolve the

 

126

past. Further, although psychotherapy in adolescence was not directly related to
adult attachment security, it was signiﬁcantly related to engaging in and the
duration of individual therapy in adulthood. larger sample sizes are needed to
map the speciﬁc trajectories related to alternate attachment ﬁgures,
psychotherapy, and child rearing.

Finally, although maternal IQ was associated with attachment security,
resolved mothers were not signiﬁcantly more intelligent than insecure mothers
and intelligence was not related to the type or duration of psychotherapy.
Therefore, the ability to engage in the process of resolution does not appear to be
limited by the level of intellectual functioning.

These are promising findings which highlight the powerful impact that
siblings, extended family members, teachers, and psychotherapists can have on the
course of individual lives throughout development. They also underscore the
importance of creating alternate experiences and providing substantive
intervention for families who are at risk for problematic parent-child relationships.
In addition, they are suggestive of the potential for supplemental experiences such
as, Big Brother and Big Sister programs.

While it is encouraging to know that there is the possibility for this level of
plasticity in development, there is also a significant toll associated with the work
of reparation. The resolved mothers described years of intense turmoil and
emotional pain as they struggled to come to terms with their childhood
experiences. Despite their progress in this regard, they often continued to be

haunted by feelings of doubt about their parenting skills, feelings of depression,

 

127

and a sense of alienation from others. What is heartening is that their children
have ﬂourished in spite of these concerns. Emotionally, these children were open
and engaging, interacting smoothly with their mothers, myself, and the assistants
in this study. Their attachment stories reﬂected responsive, nurturing caregiving
from parents and confident, contented children. Intellectually, they exceeded even
the children of secure mothers. Clearly, the loving experiences that the resolved
mothers were deprived of in their childhoods, they created for their own children.
With respect to parental intellectual functioning, it is interesting that
resolved mothers scored average to above average on three of the subtests of the
WAIS-R but were below average on Picture Arrangement (PA), thus depressing
their overall IQ scores compared to secure mothers. The PA subtest is comprised
of scenes depicting interpersonal exchanges. This raises the possibility that the
poor performance of resolved mothers on this task reﬂects remnants of their
emotional discomfort regarding interpersonal relationships which interfered with
their cognitive functioning. These findings pose provocative ideas about the

nature of intelligence.

Intelligence

There is no known research linking maternal attachment security to child
intellectual functioning so the relation between these variables in this study is an
intriguing ﬁnding which suggests that psychological and social variables play a
significant role in intellectual development. Theoretically, this is consistent with

the underlying assumption of attachment theory which posits that as children

128

develop a secure attachment relationship with their caregiver(s). the experience of
this ”safe base” creates a context for exploration and learning. The data from this
study support this position. Children of resolved and children of secure mothers
consistently performed higher across multiple areas of cognitive functioning
compared to children of insecure mothers. In particular, children of resolved
mothers scored 22 points higher on overall IQ scores compared to children of
insecure mothers, despite the fact that resolved and insecure mothers were within
the Average range of intellectual functioning. These differences among children
of resolved, secure, and insecure mothers could not be completely accounted for
by parental IQ, education, age or SES. They were uniquely related to the quality
of parent-child interactions. In particular, child IQ was strongly related to the
degree of parent-child synchrony, independent of the inﬂuence of maternal IQ.
Collectively, these results provide empirical evidence for the impact of parent-
child interactions on childrens’ cognitive development. Whether these differences
will continue to be present at later points in development is not known but it is of

major significance that they are apparent at this juncture.

Limitations of the Study and Directions for Future Research

The limitations and cautions posed by the characteristics and size of the
sample in this study cannot be underestimated. The subjects were a very
homogeneous group of Caucasian, middle-class, married, well-educated, intelligent
women in their thirties. This presents several interesting considerations. On the

one hand, it is remarkable that differences in parenting behavior and child

 

129

outcome emerged as strongly as they did given the size and similarity of the
subject pool. Further, it provides evidence that these differences are not due to
the inﬂuences associated with age, intelligence, education, socioeconomic, or
marital status. On the other hand, it is not known to what extent these results
can be generalized to samples varying on these dimensions. In addition, the mall
sample size precluded path analyses that would have helped clarify mediating and
moderating effects and interactions among the variables delineated in the
theoretical model.

From a conceptual point of view, there are many remaining questions. For
instance, the current study does not address the role of signiﬁcant adult
attachments and relationships in the reorganizational process. As Quinton &
Rutter (1988) and Ricks (1985) have discussed, spouses and significant
relationships in adulthood may be crucial in redirecting the course of one’s life.
Thus, it is not known how adult relationship experiences may be related to
alternate attachment figures at earlier points in development and psychotherapy.
Similarly, it is not known to what extent fathers may play a mediating role in child
outcome. There is no reason to believe that the mother-child relationship is the
sole source of child attachment security, particularly in light of changes in the
nuclear family whereby fathers and daycare workers are consistently providing
more child care. Finally, the patterns delineated in the current study provide data
about parent-child relationships at a particular point in time. It will be necessary
to follow how these patterns track over time to identify the variables and

experiences associated with changes in developmental trajectories.

 

130
Despite these limitations, the current study adds to the growing body of

literature pertaining to attachment security and the replication of problematic
parent-child relationships across generations. It points to core experiences
underlying emotional and psychological recovery and resiliency. As such, it holds
promise for future researchers, clinicians, and parents who are invested in

creating a healthier emotional climate for children and families.

 

APPENDIX A

 

APPENDIX A
Adult Attachment Questionnaire
Starting from as far back as you can remember, please describe your

relationships with both of your parents. (Please try to provide speciﬁc
examples).

If you had to choose four adjectives that reﬂect your childhood relationship
with your mother, which four would you choose?

 

 

 

Now, for ea_ch_ adjective say why you think that particular adjective is a
good descriptor of your childhood relationship with your mother. Why did
you choose that adjective? (Try to give a speciﬁc example of a time when
she did or said a certain thing or acted in a particular way. Remember
that detail is important).

First adjective:

Second adjective:

Third adjective:

Fourth adjective:

If you had to pick four adjectives to describe the most important emotional
reactions you had to your mother (the most important ways she made you
feel or affected your feelings), which four would you choose?

 

 

 

131

132

If you had to choose four adjectives that reﬂect your childhood relationship
with your father, which four would you choose?

 

 

Now for each adjective, please describe why you think that particular
adjective is a good descriptor of your childhood relationship with your
father. Why did you choose that adjective? (Please give a specific example
of a time when he did or said a certain thing or acted in a particular way.
Remember that detail is important).

First adjective:
Second adjective:
Third adjective:

Fourth adjective:

If you had to pick four adjectives to describe the most important emotional
reactions you had to your father (the most important ways he made you
feel or affected your feelings), which four would you choose?

 

 

During your childhood, it is very likely that you felt closer to either your
mother or your father. To which parent did you feel the closest to and
why? Why wasn’t there this feeling with the other parent?

The next three questions relate to how you got comforted as a child.
Searching back in your memory as far as you can, please describe in detail
what you would typically do if you were in need of comforting. For each
question, please say whether you would tell anyone, and if you did, whom
did you tell, and how would that person typically respond (e.g., hold you,
talk to you, tell you a joke, play a game, ignore you, tease, you, etc)?

9a.

9b.

9c.

10.

11.

12.

13.

133

When you were emotionally upset as a child what would you do? Where
there typical events or things that made you emotionally upset (e.g., hurt
feelings, disappointments, bad dreams, being scared, problems with friends,
etc)? Please describe a speciﬁc incident or example that you remember. If
you did turn to your parents when you were upset, how did they respond?

What would happen when you were physically hurt or injured? Please
describe a speciﬁc example (e.g., What happened? Did you tell anyone?
Whom did you tell? How did they respond)?

Finally, what would happen when you were ill as a child? Again, please
provide a speciﬁc example. How did your parents respond?

Please describe the first time you remember being separated from your
parents. How old were you? What was the reason for the separation?
How did they respond? How did you feel about it at the time? Are there
any other separations that stand out in your mind?

Did you ever feel rejected as a child? Looking back on it now, you may
realize it was not really rejection but as a child, do you remember feeling
rejected? How often did you feel rejected? What were the circumstances?
How old were you and what did you do?

When you were a child, were your parents ever threatening to you in any
way -- maybe because they were angry, or for discipline, or maybe just
jokingly? For example, did your parents ever threaten to hurt you, to leave
you, or send you away?

Did they ever call you names and/or say that you were a bad child? Did
this happen frequently?

How did you go about getting attention as a child? How did you get your
parents respond to and/or attend to you and your needs? Please provide a
specific example.

 

 

14.

15.

16.

17.

18.

19.

20.

134

Did your parents take an interest in your activities and accomplishments?
How much importance was placed on your activities and achievements?

Did your parents seem to take more of an interest in your activities and

accomplishments than in your emotional needs or was there a balance
between the two?

At this point, you have probably shared a lot of childhood memories
regarding your relationships with your parents. In your opinion, why did
your parents behave the way they did toward you when you were a child?

What effect, if any, did your early experiences and relationships with your
parents have on your adult personality?
On your adult relationships with friends and lovers?

What effect, if any, did your early experiences and relationships with your
parents have on your relationship(s) with your child(ren)?

Have there been many changes in your relationships with your parents
since childhood?

In what ways have these relationships changed over the years?

In what ways have they remained the same?

Describe your current relationship with each of your parents. What is the
relationship like for you now as an adult?

 

For most of us, there were some aspects of our childhood relationships

with our parents that were, to varying degrees, troublesome, conﬂictual

and/or problematic. Do these issues remain unresolved and still active?
Or, have these troublesome aspects of your childhood relationships with
your parents been resolved and/or put to rest?

Please describe in detail the process of how you have dealt with these
troublesome or problematic aspects of your childhood relationships with
your parents.

 

21.

22.

23b.

23c.

23d.

24.

135

Is there any particular thing which you feel you have learned above all
from your own childhood experiences?

What would you hope your child might have learned (or will learn) from
his/her experience of being parented?

How do you respond now, in terms of feelings, when you separate from
your own child?

The next four questions relate to loss of a parent or a signiﬁcant other. If
this does not pertain to you, please skip to question 24.

Did you experience the loss of a parent or other close loved one (sibling or
close family member) while you were a young child?

Please the describe the circumstances (e.g., How old were you? How did
the death occur? How did you respond? Did you attend the funeral?)

Have your feelings regarding this death changed much over time? If so,
please state how.

Have you had any other important loses in your adulthood? If so, please
describe using the same questions as in number 24a.

How have these losses affected your adult personality? Do they affect your
approach to your own child?

Is there anything else you would like to add?

If for any reason you feel upset or unsettled and would like to talk about your
reactions to these questions, please feel free to call me at (517) 337-0581. If I am
not available, please leave your name and phone number and I will return your
call as soon as possible.

 

136

If I need you to participate in Phase II of this study, I will call you to arrange an
appointment. If you are not called to participate in Phase II, I will still keep your
name and address on ﬁle and will send you the results of this study once it is
completed. It will probably be sometime between the Fall of 1993 and the Spring
of 1994.

I would like to thank you for your time and thoughtful effort in completing these
questionnaires. Through the contribution of individuals like yourself, who
volunteer their time and share part of their personal experiences, we are able to
gain a greater understanding of family relationships and the impact they have on
our lives. Thank you.

 

 

APPENDIX B

 

 

APPENDIX B

Attachment Relationships Interview

Childhood

1.

I’d like you to think back to early childhood. When you were upset,
in trouble or needing help, who did you turn to? (If subject says no

one, proceed with questions a, b and c. Otherwise, go to question
#2).

a. Can you give me a couple of examples of times when you handled
something yourself?
b. Why do you think it was that there wasn’t anyone for you?

c. What do you think you learned from this?

a. How did s/he usually respond?

b. Can you give me a specific example?

Why did you choose this person to turn to?
What did you do if s/he was not available?
Can you give me a specific example?
When and how did this relationship begin?
How did it change over time?

How do you account for those changes?
When and how did this relationship end?
How did you feel about it ending?

How did this relationship compare to the relationship you had with either
of your parents?

What did/have you learn(ed) from this relationship?

137

 

138

Adolescence

1. Now I’d like you to think back to your adolescent years. When you
were upset, in trouble or needing help, who did you turn to? (If
same person from childhood, go on to adulthood. If still no one
there to turn to, ask for a couple of examples of handling something
herself, why this was and what she learned).

2. How did s/he usually respond?

Can you give me a speciﬁc example?

3. Why did you choose this person to turn to?

4. What did you do if s/he was not available?

Can you give me a specific example?

5. When and how did this relationship begin?
How did it change over time?

How do you account for those changes?

6. When and how did this relationship end?

How did you feel about it ending?

7. How did this relationship compare to the relationship you had with either
of your parents?

8. What did/have you learn(ed) from this relationship?

 

139

Adulthood

1. Now I’d like you to describe the closest relationship you’ve ever had in
your adult life.

2. What is/was it about this relationship that allows(ed) you to be as close as
you are/were to this person?

3. Do you wish it were less or more intimate? Why?

What prevents or prevented this from happening?

4. What is/was the most fulﬁlling aspect of this relationship?

What is/was the most disappointing?

5. When and how did this relationship begin?
How did it change over time?

How do you account for those changes?

6. When and how did this relationship end?

How did you feel about it ending?

7. How does/did this relationship compare to other relationships you have
had in the past?

8. How did/does it affect your relationships with other people?

9. What did/have you learn(ed) from this relationship?

140

(himnt Life

1.

Finally, I’d like you to describe your closest relationship at this point in
your life. (If same person from adulthood, ask subject to describe second
closest relationship in their current life).

What is it about this relationship that allows you to be as close as you are
to this person?

Do you wish it were less or more intimate? Why?

What prevents this from happening?

What is the most fulﬁlling aspect of this relationship?

What is the most disappointing?

When and how did this relationship begin?

Do you think it will continue? Why or why not?

Has this relationship changed over time?

How do you account for those changes?

How does this relationship compare to other relationships you have had in
the past?

How does it affect your relationships with other people?

What have you learned from this relationship?

 

141

Concluding Questions

1. Have you ever been involved in any counseling or psychotherapy? If yes,
can you tell me when and what you learned from this experience?

2. Do you have any questions for me about this study or your participation in
this study?

 

LIST OF REFERENCES

LIST OF REFERENCES

Abidin, R. R. (1986). Paghtihg §treg Index. Charlottesville, VA: Pediatric
Psychology Press.

Adams, J. (1992). The relatiopship between parental perceptioh of child behavior

ahd parental childhm memories: A comparison of mothers of cgpgpct-
disturbed and non-conduct disturbed bog. Unpublished doctoral

dissertation, Michigan State University, E. Lansing.

Ainsworth, M. D., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of
attachment. New York: Basic Books.

Bakermans-Kranenburg, M. J. & van Ijzendoorn, M. H. (1993). A psychometric
study of the adult attachment interview: Reliability and discriminant validity.

Developmental chhology, 22, 870-879.

Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective.
Journal of Social and Personal Relationship; 1, 147-178.

Bartholomew, K., & Horowitz, L. M., (1991). Attachment styles among young
adults: A test of a four category model. Journal of Personality and

Sgial P§ychology, 6;, 226-244.

Belsky, J., & Isabella, R. (1988). Maternal, infant, and social-contextual
determinants of attachment security. In J. Belsky & T. Nezworski (Eds.),
Clinical implications of attachment (pp. 41-94). New Jersey: Erlbaum.

Belsky, J., Rovine, M., & Taylor, D. G. (1984). The Pennsylvania infant and
family development project, III: The origins of individual differences in
infant-mother attachment: Maternal and infant contributions. Child

Development, _S_S, 718-728.

Belsky. J ., Youngblade, L., Rovine, M., & Volling, B. A. (1991). Patterns of
marital change and parent-child interaction. Journal of Marriage and the
Family, 5_3_, 487-498.

Benn, R. K. (1986). Factors promoting secure attachment relationships between
employed mothers and their sons. Child Development, _S_7, 1224-1231.

142

 

143

Benoit, D., Vidovic, D., & Roman, J. (1991, April). Trahsmissioh of attachment
e '0 Paper presented at the Biennial Meeting of the
Society for Research in Child Development, Seattle, WA.

Bowlby, J. ( 1973). Attachment and 19g}: Vol. 2 ggpagtioh. New York: Basic
Books.

Bowlby, J. (1980). Attachment and loss: Vol. 3 !9_s§, sadhess and depression.
New York: Basic Books.

Bowlby, J. (1969/1982). Attachment and loss: Vol. 1 attachment. New York:
Basic Books.

Bretherton, I. (1994). Personal communication.

Bretherton, I., Ridgeway, D., & Cassidy, J. (1990). Assessing internal working
models of the attachment relationship: An attachment story completion task
for 3-year-olds. In M. T. Greenberg, D. Cicchetti and E. M. Cummings

(Eds.), Attachment in the preschool yeag, (pp. 273-308). University of
Chicago Press: Chicago.

Cassidy, J. (1988). The self as related to child-mother attachment at six. Child
Developmeng, §_9_, 121-134.

Cicchetti, D., Cummings, E. M., Greenberg, M. T., & Marvin, R. S. (1990).
An organizational perspective on attachment beyond infancy: Implications
for theory, measurement, and research. In M. T. Greenberg, D. Cicchetti

and E. M. Cummings (Eds.), Attachment in the preschool years, (pp. 3-49).
University of Chicago Press: Chicago.

Cohn, A. A., Silver, D. H., Cowan P. A., Cowan, C. P., & Pearson, J. L. (1991,

April). Working models of childhood attachment and marital relationships.
Paper presented at the Biennial Meeting of the Society for Research in

Child Development, Seattle, WA.

Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and

relationship quality in dating couples. Journal of Personality and Social
chhology, 3, 644-663.

Crandell, L. E. (1991). The attachment relationships interview. Unpublished
manuscript, Michigan State University, E. Lansing.

Crandell, L. E. (1992, May). Continuities and discontinuities in attachment

relationships aczoss the life span: A qualitative study. Paper presented at
the Midwestern Psychological Association, Chicago, IL.

 

144

Crowell, J. A., & Feldman, S. S. (1989). Assessment of mothers’ working models
of relationships: Some clinical implications. WWI,
19, 173- 184.

Egeland, B., & Farber, E. A. (1984). Infant-mother attachment: Factors related
to its development and changes over time. W 5;, 753-771.

Egeland, B., Jacobvitz, D. & Sroufe, L. A. (1988). Breaking the cycle of abuse.
Child ngelppment, 52, 1080-1088.

Erickson, M. F., Sroufe, L. A. & Egeland, B. (1985). The relationship between
quality of attachment and behavior problems in preschool in a high-risk
sample. In I. Bretherton & E. Waters (Eds). Growing Points of
Attachment Theom and Research, Monograph of the Socieg for Beseaich
in Child Development, 59, 147-165.

Erikson, E. H. (1959/1963). Childhood and Society. New York: W.W. Norton.
Co.

Emde, R. N. (1983). The prerepresentational self and its affective core. flfhe
Esyghoahalﬁic Study of the Child, .38, 165-192.

Feeney, J. A., & Noller, P. (1990). Attachment style as a predictor of adult

romantic relationships. Journal of Personality and Social Pachology,
5_8, 281-291.

Fonagy, P. (1993, March). Reﬂective self function in early attachment and
borderline states. Paper presented at the Biennial Meeting of the Society

for Research in Child Development, New Orleans, LA.

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 Developmenl, _6_2_, 891-905.

Fonagy, P., Steele, M., Steele, H., Moran, G. S., & Higgitt, A. C. (1991). The
capacity for understanding mental states: The reﬂective self/parent in
mother and child and its significance for security of attachment. Infant

WM]. .12. 201-218.

Fraiberg, S., Adelson, E., & Shapiro, V. (1975). Ghosts in the nursery: A
psychoanalytic approach to the problems of impaired infant-mother

relationships. J oumal of the American Academy of Child Psychiat_ry,
14, 387-422.

 

 

145

Franz. C. E., & White, K. M. (1985). Individuation and attachment in personality

development: Extending Erikson’s theory. Journal of Persohality, m, 224-
256.

Frommer, E. E., & O’Shea, G. (1973). The importance of childhood experience
in related to problems in marriage and family-building. British Journal

Mm 123.. 157-160.

George, C., Kaplan, N ., & Main, M. (1985). Adult amphment intem'ew.
Unpublished manuscript, University of California, Berkeley.

Goldberg, S., MacKay-Soroka, S, & Rochester, M. (1993, March). Mia—31.. F

attachment and matemal tespghsiveneg. Paper presented at the Biennial
Meeting of the Society for Research in Child Development, New Orleans,

LA.

 

Goldstein, W. N. (1991). Clarification of projective identification. m ‘
Jgumal of Psychiatry, 118(2), 153-161.

Grossman, K. E., Loher, 1., Grossman, K., Scheuerer-Englisch, H., Schildback,
B., Spangler, G., Wensauer, M., & Zimmerman, P. (1993, March). _Thp
development of inner working models of attachment and adaptation. Paper
presented at the Biennial Meeting of the Society for Research in Child
Development, New Orleans, LA.

Haft, W. L., & Slade, A. (1989). Affect attunement and maternal attachment:
A pilot study. Infant Mental Health Journal m, 157-172.

 

Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment

process. Journal of Personality and Social Psychology, 53(3), 511-524.
Homey, K., (1950). Neurosis and human growth. New York: W.W. Norton Co.

Isabella, R. A., & Belsky, J. (1991). Interactional synchrony and the origins of
infant-mother attachment: A replication study. Child Development, 6_2_,
373-384.

Izard, C. E., Haynes, O. M., Chisholm, G., & Baak, K. (1991). Emotional
determinants of infant-mother attachment. Child Development, Q, 906-917.

Jacobvitz, D., Morgan E., Kretchmar, M., & Morgan, Y. (1991, April). Parenting
acrpss three generations: The development of maternal empathy. Paper
presented at the Biennial Meeting of the Society for Research in Child
Development, Seattle, WA.

146

Kobak, R. R., & Hazan, C. (1991). Attachment m marriage: Effects of security

and accuracy of working models WW
_wghglgu. 60 861-869

Kobak, R. R. & Sceery, A. (1988). Attachment in late adolescence: Working
models, affect regulation, and representations of self and others. Child

Mm; 52. 135-146-

Kaufman, J ., & Zigler, E., (1987). Do abused children become abusive parents?
Americah Qﬂhgpgchiatric Amocjation, 5_7(2), 186-192.

Lamb, M. E. & Easterbrooks, M. A. (1981). Individual differences in parental P
sensitivity: Origins, components, and consequences. In M. Lamb & L.
Sherrod (Eds.), Infant and social co nitio : E i 'cal a theo etica
considerations (pp. 128-153). New Jersey: Erlbaum.

 

Main, M. (1981). Avoidance in the service of attachment: A working paper. In
K. Immelman, G. Barlow, M. Main & L. Petrinovitch (Eds.), Behavioral

develppment: The Bielefeld interdimiplinam project (pp. 651-693). New
York: Cambridge University Press.

Main, M. & Goldwyn, R. (1984). Predicting rejection of her infant from mother’s
representation of her own experience: Implications for the abuse-abusing

intergenerational cycle. Child Development, _8_, 203-217.

Main, M. & Goldwyn, R. (1989). Adult attachment classification system.
Unpublished manuscript, University of California, Berkeley.

Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood,
and adulthood: A move to the level of representation. In I. Bretherton

& E. Waters (Eds). Growing Eoints of Attachment Theom and Research,
Monograph of the Socieg for Research in Child Development, 59, 66-106.

Main M., & Weston, D. R. (1981). The quality of the toddler’s relationship to
mother and to father: Related to conﬂict behavior and the readiness to
establish new relationships. Child Development, 5_2, 932-940.

Murray, L. (1991). Intersubjectivity, object relations theory, and empirical

evidence from mother-infant interactions. Infant Mental Health Journal, _l_2_,
219-232.

 

147
Osofsky J D (1993 March)- MW

 

Paper presented at the Biennial Meeting of the Society for Research m
Child Development, New Orleans, LA.

Pastor, D. L. (1981). The quality of mother-infant attachment and its
relationship to toddlers’ initial sociability with peers. Qevelopmehthl

Emhology, 11(3), 326-335.
Pratt, M. W., Roth, D., Cohn, C., Cowan, P. A., & Cowan, C. P. (1991, April).
Co tin en tion : du ts’ com lexi o sonin about ami

oijgip, attachment security, and quality of parenting. Paper presented at the
Biennial Meeting of the Society for Research in Child Development, Seattle,

WA.

 

Quinton, D., & Rutter, M. (1988). Parenting breakdown. Brookﬁeld, USA:
Avebury Gower.

Ricks, M.H., (1985). The social transmission of parental behavior: attachment
across generations. In I. Bretherton & E. Waters (Eds). Growing Points of

Attachment Theoiy and Research, Monograph of the Socieg for Research
in Child Developmeht, ﬂ), 211-230.

Robinson, E. A., & Eyberg, S. M. (1981). Dyadic parent-child interaction coding
system: Standardization and validation. Journal of Consulting and Clinical
chhology, 32(2), 471-481.

Schneider-Rosen, K. (1990). The developmental reorganization of attachment
relationships: Guidelines for classification beyond infancy. In M. T.
Greenberg, D. Cicchetti and E. M. Cummings (Eds.), Attachment in the
preschool yeats, (pp. 185-220). University of Chicago Press: Chicago.

Silverstein, A. B. (1982). Two- and four-subtest short forms of the Wechsler
Adult Intelligence Scale-Revised. Journal of Consulting and Clinical
Psychology, $6), 415-418.

Spieker, S. J. & Booth, C. L. (1988). Maternal antecedents of attachment quality.

In J. Belsky & T. Nezworski (Eds.) Clihical implications of attachment (pp.
95-135). New Jersey: Erlbaum.

Sroufe, L. A. (1983). Infant-caregiver attachment and patterns of adaptation
in preschool: The roots of maladaptation and competence. In M.

Perlmutter (Ed.), Minnesota Smposium in Child Psychology, 1_6, 41-81.

148

Sroufe, L. A. (1988). The role of infant-caregiver attachment rn development.

InJ. Belsky&T. Nezworski (Eds) WW6?
18-38). New Jersey: Erlbaum.

Sroufe, L.A., (1989). Relationships, self, and individual adaptation. In

A.J. Sameroff & R.N. Emde (Eds). Relationship disturbances in early
2111mm, (pp. 70-94). New York: Basic Books.

Sroufe, L. A., & Fleeson, J. (1986). Attachment and the construction of
relationships. In W. W. Hartup and Z. Rubin (Eds.), Relationships and

Development (pp. 51-71).

Sroufe, L. A., Pancake, V. R., & Fox, N. E. (1983). Attachment and dependency
in developmental perspective. Child Develgpment, 51, 1615-1627.

Stern, D. N. (1985). The intei‘personal world of the infant. New York: Basic
Books.

Stevens, G. & Featherman, D. L. (1981). A revised socio-economic index of
occupational status. Social Science Research _1(_), 364-395.

 

Thompson, R. A., Lamb, M. E., & Estes, D. (1982). Stability of infant-mother
attachment and its relationship to changing life circumstances in an
unselected middle-class sample. Child Development, ﬂ, 144-148.

Thorndike, R. L, Hagen, E. P., & Sattler, J. M. (1986). The Stanford-Binet
Intelligence Scale. Chicago: The Riverside Publishing Company.

Tracy, R. L., & Ainsworth, M. D. S. (1981). Maternal affectionate behavior and
infant-mother attachment patterns. Child Development, _5_2_, 1341-1343.

Tronick, E. Z., Ricks, M., & Cohn, J. F. (1982). Maternal and infant affective
exchange: Patterns of adaptation. In T. Field & A. Fogel (Eds.), Emotion
and early interaction (pp. 83-100). New Jersey: Erlbaum.

 

Vaughan, B., Egeland, B., & Sroufe, LA. (1979). Individual differences in
infant-mother attachment at twelve and eighteen months: Stability and
change in families under stress. Child Development, 4_9_, 483-494.

Wamboldt, F. S., (1991, April). Transmission of relational patterns from origin
family to early marriage. Paper presented at the Biennial Meeting of the

Society for Research in Child Development. Seattle, Washington.

Waters, E., (1978). The reliability and stability of individual differences
in infant-mother attachment. Child Development, 4_9_, 483-494.

149

Wechsler, D. (1981). jlhg Wechglet Adplt lptelliggpgg S_«mle - Revised.
Cleveland: The Psychological Corporation.

Weiss, R. S., (1982). Attachment in adult life. In C. Parkes & J. Stevenson-

Hinde (Eds.), flhe place of attachment in human behavior, (pp. 171-184).
New York: Basic Books.

Whipple, E., Fitzgerald, H. E. & Zucker, R. (In Press). Parent-child interactions
in alcoholic and non-alcoholic families. Amen’gp ,loumal of

Orthgpgchiatry.
Whipple, E., Denberg, N. L., & Davies, W. H. (1993). MSU family prpject
e 'si 0 els e -c ' ° teraction codin te .

Unpublished manuscript, Michigan State University, E. Lansing.

Zucker, R. A. & Fitzgerald, H. E. (1991). Risk and coping in children of

alcoholics: Years 6 to 10 of the Michigan State University longitudinal
study. NIAAA Grant Proposal.