Mac Arthur Story Stem Battery

THE REFLECTION
OF PATTERNS OF ATTACHMENT
IN NARRATIVES
OF PRESCHOOL
IN INFANCY
CHILDREN
Susan Anne Yabsley
Submitted for the degree of Doctor of Philosophy

University College London
1999
BML
(LNDON
(D,
UNIV
2
ACKNOWLEDGEMENTS
I am very grateful to many people for their invaluable support and help throughout
the course of this Ph.D. To my supervisor, Peter Fonagy, I wish to express my heartfelt
thanks for his patience, understanding and guidance. To Drs. Howard and Miriam Steele,
firstly for inviting me to join the London Parent-Child Project, but mostly for believing
in me and offering me a truly special friendship. It has been a privilege and a pleasure to
work with all three.
I also wish to thank my co-workers on PhaseThree of the study, Dr. Juliet Holder,
Dr. Matthew Woolgar, Dr. Carla Croft, Dr. Jo Alves and Dr. Jacques China. It was the
in
friendly
I
team
that
such a
and supportive
experience of working
will remember
fondly. I also owe thanks to all my friends and colleagues at The Anna Freud Centre.
Finally, I would like to thank my parents for making all of this possible. I could
it
have
accomplished without them.
not
-I,
ABSTRACT
This series of studies reports on the prospective and concurrent relation of

internal
five
It
based
the
to
world.
attachment a narrative
year old child's
assessmentof
designed
to explore the
the
relies on
exploration of a relatively new research measure
have
investigated
No
the validity
the
published studies
yet
internal world of
young child.
in
in
this
to
thoroughly
tested
of
measuresof attachment patterns infants
measure relation
investigate
independent
Additionally,
this
the
study will
contributions of
and parents.
father.
mother and
The first two chapters review the literature and introduce the instruments to be
initial
internal
The
the
theoretical
the
chapter examines
points of view regarding
used.
from
the perspective of psychoanalytic, cognitive and
world and mental representations
discusses
level
in
It
the
theory.
then
to
move a
of representation attachment
attachment
has
Chapter
history
Two
that
the
the
current
made
study possible.
considers
research
of
the technique of doll play as a researchtool and examines the scant researchthat has been
Story
Stem
Battery.
MacArthur
the
published using
Chapter Three responds to a need for psychometric information regarding the
MacArthur Story Stem Battery and the corresponding MacArthur Narrative Coding
System by reporting on the construction of reliable and valid factors/scales. Subsequent
demographic
before
to
these
the
scales associations
variables collected
chapters present
birth and during the infancy of the target children. Later chapters report on the
longitudinal and concurrent associations between the scales with categories of infantparent, child-parent and parental representationsof attachment security and with parental
behaviours
utilizing the reliable and validated Child
assessments of child problem
Behaviour Checklist (CBCL). In addition, an attempt is made to construct theoretical
insecure
battery
to
the
children's responses
story stem
profiles of secure and
and to apply
these profiles to the four groups of attachment patterns in the sample.
The discussion focuses upon the creation of psychometrically valid scalesrelevant
to important aspectsof the child's internal world. It also concentratesupon discussing the
low-risk,
the
these
to
this
results
of
application
of
scales
confirmed and unconfirmed
Project.
London
Parent-Child
the
non-clinical sample of
4
TABLE OF CONTENTS
Chapter 1
15
The Representational World: Theory and Research
...........................
16
1.1 Introduction
................................................................................................
16
1.1.1 A Psychoanalytic Approach
........................................................
1.1.2 A Cognitive Approach
.................................................................-13
1.2 Attachment Theory and the Internal Working Model
27
...............................
1.2.1 Assessment of attachment in infants
31
..........................................
1.2.2 A move to assessingthe representational level
35
..........................
1.2.3 Recent perspectives on attachment theory
38
.................................
1.3 Conclusions
43
...............................................................................................
Chapter 2
2.1
2.2
2.3
2.4
Assessing the Internal World of the Young Child

44
................................
Introduction
45
...............................................................................................
Doll play techniques
46
..................................................................................
Child psychoanalysis
46
.................................................................................
Assessing internal working models of young children
48
.............................
2.4.1 Representation and narrative
49
......................................................
2.4.2 The MacArthur Story Stem technique
53
.......................................
2.4.3 The MacArthur Narrative Coding System
56
.................................
2.5 The London Parent-Child Project
57
.............................................................
2.6 Conclusions
59
...............................................................................................
Chapter 3
3.1
3.2
3.3
3.4
3.5
Analysis of the MacArthur Story Stem Battery Narratives

Introduction
62
..............................................................................................
Method
63
.....................................................................................................
Results
71
......................................................................................................
Discussion of reliability results
81
................................................................
Reduction of data
87
.....................................................................................
3.5.1 Scaling
87
......................................................................................
3.5.2 Factor analysis
90
..........................................................................
Chapter 4
The Effects of Demographic Factors on the Children's Responses to

Narrative
Task
99
a
...................................................................................
Introduction
100
.............................................................................................
Method
106
.....................................................................................................
Results
110
.....................................................................................................
4.3.1 Age
1
...........................................................................................
4.3.2 Expressive language and level of education
I
............................
4.3.3 Infant temperament
113
..................................................................
4.33.4 Social class
114
...............................................................................
4-3.5 Gender
116
......................................................................................
Discussion
122
...............................................................................................
4.1
4.2
4.3
4.4
5
Chapter 5
5.1
5.2
5.3
5.4
Chapter 6
6.1
6.2
6.3
6.4
Chapter 7
7.1
7.2
7.3
7.4
Associations between the Child Behaviour Checklist and the

127
MacArthur Story StemBattery
...........................................................
128
Introduction
.............................................................................................
1 2
Method
..................................................................................................... -3)
134
Results
.....................................................................................................
134
5.3.1 Agreement between parents
.....................................................
136
5.3.2 Mother's CBCL rating
..............................................................
137
5.3.3 Post hoc analysis
.......................................................................
141
5.3.4 Father's CBCL rating
...............................................................
142
.......................................................................
143
Discussion
...............................................................................................
Associations Between the MacArthur Story Stem Battery and ChildParent Attachment
148
...............................................................................
Introduction
149
.............................................................................................
Method
152
....................................................................................................
Results
156
.....................................................................................................
6.3.1 Child-mother data
156
....................................................................
6.3.2 Child-father data
157
......................................................................
Discussion
158
...............................................................................................
Associations Between the MacArthur Story Stem Battery and Infant162
Parent Attachment
...............................................................................
163
Introduction
.............................................................................................
167
Method
.....................................................................................................
169
Results
.....................................................................................................
171
7.3.1 Infant-mother data and the story stem factors
..........................
7.3.2 Infant-mother data per story
176
.....................................................
7.3.3 Infant-mother data per variable
179
................................................
7.3.4 Infant-father data and the story stem factors
182
............................
7.3.5 Infant-father data per story
185
.......................................................
7.3.6 Infant-father data per variable
186
..................................................
Discussion
189
...............................................................................................
Associations Between the MacArthur Story Stem Battery and

Chapter 8
196
Parents' Attachment Classification
.................................................................
8.1 Introduction
197
.............................................................................................
8.2 Method
200
....................................................................................................
8.3 Results
203
.....................................................................................................
8.3.1 Mother data
204
..............................................................................
8.3.2 Post hoc analyses
206
......................................................................
8.3.') Father data
207
................................................................................
83.4 Post hoc analyses
209
......................................................................
8.4 Discussion
211
...............................................................................................
6
Chapter 9
Interaction of Gender and Attachment ...............................................214
215
9.1 Introduction
...................................................................................... ***....
218
9.2 Method
....................................................................................................
219
9.3 Results
..................................................................................................... 219
9.3.1 Gender and infant-mother attachment

......................................
220
......................................................................
222
9.3.3 Gender and infant-father attachment
........................................
2233
......................................................................
9.4 Discussion
...............................................................................................
227
230
Chapter 10 Profile Analysis
...................................................................................
231
10.1 Introduction
...........................................................................................
233
10.2 Method
...................................................................................................
2333
10.2.1 Sample and design
..................................................................
234
10.2.2 Procedure - Development of the Profile
.................................
10.2.3 Prototypical profiles
235
...............................................................
237
10.3 Results
...................................................................................................
10.3.1 Inter-rater reliability
237
...............................................................
241
10.3.2 Profile analysis
.......................................................................
242
10.4 Discussion
.............................................................................................
Chapter 11
11.1
11.2
11.3
11.4
247
Multivariate Regression Analysis
.......................................................
248
Introduction
...........................................................................................
248
Method
..................................................................................................
249
Results
...................................................................................................
262
Discussion
.............................................................................................
Chapter 12
12.1
12.2
12.3
12.4
264
Summary and Conclusions
..................................................................
265
Introduction
...........................................................................................
266
Confirmed Findings
..............................................................................
274
Unconfirmed Findings
...........................................................................
276
Conclusions
...........................................................................................
TABLE OF APPENDICES
A. I
A. 2
A. 33
A. 4
A. 5
294
The MacArthur Story Stem Battery protocol
.................................................
300
MacArthur Narrative Coding Manual
............................................................
Coding Sheets
321
....................................................................................................
Frequency of Content and Performance Codes Observed
325
..............................
T-tests for differences in the means of the factors by story
340
............................
B. I
B. 2
B. 3
Mill Hill Vocabulary Scale

342
.............................................................................
Infant Characteristics Questionnairre
344
..............................................................
The Bus Story Test
348
..........................................................................................
C. I
Child Behaviour Checklist
D. I
D. 2
AAI Questions
352
.................................................................................................
Four way grouping of mother's attachment
353
....................................................
E. I
Interaction of gender and attachment

354
..............................................................
F. I
Coding Sheet for Profile
G. I
G.2
358
Discriminant Function Analysis
......................................................................
364
Cluster analysis
................................................................................................
..............................................................................
..................................................................................
350
355
TABLE
OF TABLES
61
.........................................................................................
Table 2.1
Project Design
Table 3.1
Median Cohen's Kappa scores for the inter-rater reliability of the 24

in
72
themes
the
as
presented
manual
content
..........................................
Table 3.2
Median Kappa scores for inter-rater reliability of the complex content

74
themes collapsed to form binary variables
..........................................
Table 3.33
Median Kappa scoresfor inter-rater reliability of the new variables derived

from the complex content themes
74
.........................................................
Table 3.4
Median inter-rater reliability for the parental representations manual

binary
75
the
categories coding and
coding ...............................................
Table 3.5
Median inter-rater reliability for mother and father representations 75

.....
Table 3.6
Median inter-rater agreement for the performance codes as presented

in the manual
77
.........................................................................................
Table 3.7
Median scores for inter-rater reliability of the complex performance

78
themes collapsed to form new variables
................................................
Table 3.8
Median inter-rater reliability for narrative coherence

78
...........................
Table 3.9
Median inter-rater reliability for the affect themes as presented in the

manual .................................................................................................. 79
Table
10
Median inter-rater reliability of the complex affect themes collapsed to

form binary variables
80
............................................................................
Table 3). 11
Median inter-rater reliability of affect themes when three phases of the

narrative task are combined .................................................................. 81
Tab Ie
I
-'). -'
Reliable content codes and aggregate alphas across the II stories

88
......
Tab Ie
133
-').
Parental representations and their aggregate alphas across storles
.......
89
14
Reliable performance and affect codes and their aggregate alphas across
stories.................................................................................................... 90
Fable '). 15
Sorted factor loadings for the unrotated solution

91
..................................
Tab Ie3.16
Sorted factor loadings for the varimax rotation

91
.....................................
9
Table 3.17
Means, standard deviation, range and Cronbach's alpha per story by

I 92
factors one (n=89) and two (n=86)
........................................................
Table 3.18
-3
Means,, standard deviation, range and Cronbach's alpha per story by

93
factors three (n=89) and four (n=86)
....................................................
Table 3.19
Intercorrelations between the eleven story stems for factor 1, quality/open

response to task ..................................................................................... 94
Table 3.20
Presentsthe Intercorrelations between the eleven story stems for factor 2.

95
discipline/punishment
...........................................................................
Table 3.21
Presents the Intercorrelations between the eleven story stems for factor
95
three, negative/controlling
.....................................................................
Table 3.22
Presents the intercorrelations between the eleven story stems for factor
four, positive matemal. representation
96
...................................................
Table 3.23
97
The composition of the four factor scales
..............................................
Table 4.1
Bivariate correlations and levels of significance of mother's, father's and

four
factors
III
the
age
and
story
stem
children's
........................................
Table 4.2
Bivariate correlations and levels of significance for the Bus Story scales
factors
four
112
the
story stem
and
............................................................
Table 4.3
Correlations and levels of significance for parents' expressive language

four
level
factors
113
the
story stem
with
and education
...........................
Table 4.4
Correlations and levels of significance of infant temperament and the story

factors
114
stem
.........................................................................................
Table 4.5
Means and SDS of factor scores grouped by three classifications of

mother's social class............................................................................ 115
'Fable 4.6
Means and SDSs between factor scores grouped by three classifications

father's
social class......................................................................... 116
of
Table 4.7
Means, T scores and levels of significance for the Bus Story scales
by
gender............................................................................... 117
grouped
I'able 4.8
Means, SDS and levels of significance by gender for the four story stem
factors
118
..................................................................................................
10
Table 4.9
Means, SDS and levels of significance for the content and parental
by
0
122
grouped
gender
representation variables
........................................
Table 4.10
Means, SDS and levels of significance for the performance variables by

121
gender..................................................................................................
Table 5.1
Correlations of mother's and father's rating on CBCL scales

135
.............
Table 5.2
Correlations and levels of significance of the four story stem factors with
mother's rating on the CBCL .............................................................. i) ',6
Table 5.3
Correlations of mother's rating of the child on the intemalising scale with

factor
Discipline/Punishment
138
variables comprising
............................
Table 5.4
Correlations of mother's ratings of the child on the intemalising scale with

factor
Positive
Matemal
Representation
138
comprising
variables
............
Table 5.5
Correlations of mother's rating of the child on the externalising scale with

factor
ling
Negative/Control
139
...............................
Table 5.6
Correlations of physical aggressionand the extemalising scale as rated by

for
boys
140
mother
girls and
......................................................................
Table 5.7
Correlations and levels of significance of father's rating of the child on the

CBCL for the four story stem factors
141
..................................................
Table 5.8
Correlations of father's rating of the child on the extemalising scale with

factor
ling
Negative/C
142
ontrol
...............................
Table 5.9
Correlations of physical aggressionand the externallsing scale as rated by

father for girls and boys
143
.......................................................................
Table 6.1
Means and SDS of the four story stem factors grouped by child insecurity
five
years of age....................................... 157
vs. security with mother at
Table 6-2
Means and SDS of the four story stem factors grouped by child insecurity
five
half
father
158
a
at
and
years of age.......................
vs. security with
Table 7.1
Infant-parent security grouped by instrument. forced classlficatlon in

parentheses........................................................................................... 168
Table 7.2
Means and SDS of the four story stem factors grouped by infant insecurity
171
twelve
months .........................................
vs. security with mother at
Table 7.3
Means and SD of father's social class and security of attachment to mother

for
factor
12
172
the
at
months
.............................
Table 7.4
Means and SIDS of story stem factors grouped by original three-way

infant
classifications of
attachment with mother at 12 months ............ 17)
Table 7.5
Mean rank of the Discipline/Punishment factor grouped by three-way

attachment classification with mother at twelvemonths ...................... 174
Table 7.6
Means and SDS of the four story stem factors by four-way classification
infant
security with mother at 12 months ....................................... 175
of
Table 7.7
Means, SDS of story stem factors by three-way classification of

12
disorganised
(excluding
attachment with mother
months
group)... 176
Table 7.8
Means and SDS of insecure vs secure attachment with mother at 12

by
for
factor
Quality/Open
Response
177
the
months
story
......................
Table 7.9
Means and SDS of secure vs insecure attachment with mother at 12

factor
for
178
each story on
months
......................
Table 7.10
Means and SDS for the content and parental representation variables
insecure
by
vs secureattachment classification with mother at 12
grouped
179
months...................................................................................................
'Fable 7.11
Means and SDS of performance variables grouped by insecure vs secure

12
181
months ..................................................
attachment with mother at
Table 7.12
Means and SDS of the four story stem factors grouped by infant insecurity
father
182
months
at
eighteen
security
with
vs.
.........................................
Table 7.1
Means and SDS of the four story stem factors grouped by three
father
infant
18
18
3.
attachment security with
at
mths..
classifications of
Table 7.14
Means and SDS of four story stem factors grouped by avoidant and secure
father
disorganised
18
(excludes
months
children).. 184
attachment with
I'able 7.15
Means and SDS of securevs insecure attachment with father at 18 months

for each story on Quality/Open Response factor
185
.................................
Fable 7.16
Means, SDS and levels of significance for the content and parental
by
insecure vs secure attachment classification
father
18
months ...................................................................... 187
at
with
12
'Fable 7.17
Means and level of significance for the performance variables by insecure

father
18
188
to
at
months
vs secure attachment
.......................................
Table 8.1
Means and SDS of the four story stem factors with mother's AAI
(insecure
before
birth
vs
secure)
subject .....204
attachment classification
Table 8.2
Means and SD of father's social class and mother's AAI attachment

factor
205
the
classification with
..........................
Table 8.3
Means and SDS of story stem factors grouped by three-way classifications

interview
206
of mother's adult attachment
...............................................
Table 8.4
Means, SDS and levels of significance of four factors with father's AAl
(insecure
before
birth of subject..208
vs secure)
Table 8.5
Means and SD of father's social class and father's AAI attachment

classification with the Discipline/Punishment factor ...........................209
Table 8.6
Mean rank of the first two story stem factors grouped by three-way
father's
interview
210
classification of
adult attachment
...........................
Table 8.7
Means and SDS of story stem factors grouped by three-way classification

father's
interview
211
of
adult attachment
.................................................
Table 9.1
Means and SDS of the two-way interaction of gender and insecure vs

for
the four
secure attachment classification with mother at one year
factors
220
..................................................................................................
Table 9.2
Means and SDS of two-way interaction of gender and insecure vs secure

for
12
the content and
months
attachment classification with mother at
221
parental representation variables ........................................................
Table 9.33
Means and SDS of the two-way interaction of gender and insecure vs

father
18
for
four
the
at
secure attachment classification with
months
factors
222
..................................................................................................
Table 9.4
Means and SDS of two-way interaction of gender and insecure vs secure

attachment classification with father at 18 months for the content and
parental representationvariables.......................................................... 224
Table 9.5
Means and SIDSof two-way interaction of gender and insecure vs secure

father
18
classification
attachment
with
at
months with the perfon-nance
variables................................................................................................226
1-
I-,
Table 10.1
High (or present) and low (or absent) ratings per variable per story where
I
Q
-),
3
agreed
1"
all raters
..............
...................................................................
Table 10.2
Percentage of agreement for content, parental representation and

performance variables . .......................................................................240
Table 10.3
Mean correlations of z-scoresbetween the profiles and the four attachment

groups of children .................................................................................241
Table 11.1
Discriminant function analysis grouped by infant attachment security with

mother ................................................................................................251
Table 11.2
Discriminant function analysis using the factors calculated per story as

discriminating variables grouping by infant attachment security with
mother per story .................................................................................253
'Fable 11.33
Discriminant function analysis grouped by infant attachment security with

father
254
...................................................................................................
-rable 11.4
Discriminant function analysis using factors calculated per story as

discriminating variables grouped by infant attachment security
father
256
with
............................................................................................
Table 11.5
258
Discriminant function analysis grouped by mother's AAI
................
Table 11.6

259
discriminating variables grouped by mother's AAI
.............................
Table 11.7

261
discriminating variables grouped by father's AAI
...............................
14
CHAPTER ONE
THE REPRESENTATIONAL
WORLD:
THEORY AND RESEARCH
16
1.1
THE REPRESENTATIONAL
WORLD
Central to this study are the theoretical concepts of an individual's internal world
its
do
'representational
What
the
to
we
mean
when
world
representations.
we
refer
and
internal
do
by
9
In
?
What
*representations'
the
order to
child'
isation' and
mean
of
begin thinking about these complex ideas this chapter will review some of the literature
from
both
the
a psychoanalytic and a cognitive
internal world
about representationsand
theory perspective. It is important to consider both of these perspectivesas they will lead
That
this
theoretical
to
project.
construct which is an important aspect of
is.
us another
internal
the
theory's
working model.
construct of
attachment
1.1.1
A Psychoanalytic Approach
In the early writings of Freud (1895) the term representationidentifies the product
from
internal
is,
internalization,
the
that
separate
mental world
external
creation of an
of
internal
to
to
This
give
meaning
allows
us
our own
world
mental
of
an
creation
reality.
internalisation,
We
when an intersubjective relationship is
usually speak of
experiences.
less
A
psychic representation is a more or
transformed into an intra subjective one.
thing
the
of
a
meaningful
or object.
of
a
perception
within
mind
consistent reproduction
between
begins
he
to
As a t-mb\ t,
she
or
experience an awareness
inner and outer
ormvs
from
Memory
delineating
traces
these
of
self
non-self
perceptions
and scif and iioii-sclf.
development,
down
laid
these
with
maturation
and,
and
nuclei
of
self and object
arc
17
become
further
development
With
psychic representations
representationsare elaborated.
find
Many
the
aspects of
psycho-physiological self
more complex and unique.
Similarl.
the
the
self.
psychic representation of
v, all aspects of objects,
representation in
important
individual
find
that
the
to
are
animate and inanimate,
psychic representation
inner
&
Fine,
(Moore
that
person's representational world, an
world of objects
as part of
1990). It is through affectively invested interactions with the environment, and memories
infant
builds
himself,
that
the
these,
the other and their
mental representations of
of
&
Tyson,
1990).
(Tyson
interaction
Freud's focus was primarily on the product of interrialisation and his explanation
for how such a mental phenomenon may actually come about was subjected to extensive
for
development
drive
(1915)
His
the
theory
of mental
account
early
revision.
holds
begins
that
the
the
and
relationship
earliest
mother-infant
with
representations
bond which links mother to infant is libido. Infants come to love and represent mother
based on her ability to gratify their instinctual needs thereby discharging libido and
Freud
Later
(1926),
this
view and considered mother's
modified
reducing anxiety.
factor
the
the
the
of
anxiety
as
well
as
reduction
in
in
critical
as
absence
presence or
development of self and object representations.
One of the main ideas throughout Freud's theoretical constructions \vas

both
by
hypothesis
that
the
the
any
relation
experienced
on
expressed
undoubtcdl
level
duritig,
have
with
one's
own
parents
unconscious
infanc,
y
will
a
conscious and
dccisi,, c influci-ice on the quality of a child's mental representationsof those relationships
development
the
of
child's
the
personality.
as
as \vcll
18
In his paper On Narcissism: An Introduction (1914). Freud deals \1th parental

intergenerational
focusing
during
this
process,
roles
on the parental "compulsion to
hope
the
to
the
that "the child shall fulfil
child" and
ascribe every perfection
parental
those wishful dreams of the parents which they never carried out" (p.91). In a later essay,
Group Psychology and the Analysis of the Ego (192 1), Freud shows the other aspect of
this process, wherein he considers the mechanism of identification by the child, which
Freud
(p.
105).
"the
represents
earliest expression of an-emotional tie with another person"
believed that a transformation in internal mental structuring and representatiori takes
identification,
he
linked
to the psychological awarenessof
through
a
also
process
place
from
loved
objects, i. e., one's caregiving parents.
self as separate
In The New Introductory Lectures on Psychoanalysis(193 3), Freud elaboratedon

_3
has
lost
has
"If
by
transformation
occur.
one
an
object
or
the mechanism which such
may
been obliged to give it up (as children must eventually give up or separatefrom their
it
it
by
identifying
by
oneself with and setting up
parents), one often compensatesoneself
foreshadows
Freud
"
In
527).
(p.
this
what will
statement
once more in the ego...
how
developmental
become
research:
a major area of current
eveiitually
mental
transmitted
those
the
are
parents
with
relationships
and
representationsof ones' parents
oi- pssed on to succeeding generations.
19
Contemporary psychoanalysis
Contemporary psychoanalytic theories are characterised by a shift in the

isation
from
the
conceptual
of
unconscious
a repository of repressedinstinctual wishes to
a structure comprising representations of self, object and of prototypic interactions
between the two (Eagle, 1995). The interest in representational structures that evolve
from the process of internalisation is also an expression of a shift in interests in
from
focused
to
two
psychoanalysis
a one person psychology a
person psychology
on the
fundamental
1995).
(Blatt,
caring experiences
quality of early
Psychoanalystsfrom the independent group are likely to hold the classical view
from
But
drives
functioning
body
they also often
the
arising
it.
and of
of
importance of
for
both
form
desires
the
to
of
and about objects,
in
assign equal motivational strength
their
things
and
associated phantasies
structures
as
and
internal
and
people
external
(Rayner, 1991). One independentpsychoanalyst,Fairbaim, felt that the origins of much
from
These
in
lay
seen
as
arising
are
often
object
relations.
external
real
pathology
When
intimacy
loss
forrns
working
with parents - particularly with mother.
of
of
various
form
'emotional
have
likely
intimacy
to
or
special
own
its
seems
earIN'
SLIch
\\-ell,
trauma.
loss
delight'and
a
constitutes
the
an
experience
such
of
assurance
of
patterning of
Fairbairn
from
distortions
from
overstimulation.
Traurnata may also arise
of reality and
for
his
lost
in
individual
thenceforth searches. one way or another,
believed that the
being
'object-seeking'
human
he
In
the
rather
this
as
sees
way
substitutes.
its
or
intimacy
is
by
he
By
the
this
that
quality of
gained
pleasure
means
pleaSUre-seeking.
than simpk
discharge
than
of energy.
a I
the state of an ego-oject relation - internal or external - rather
20
Likewise anxiety is reduced by a change in object relation rather in dischar(-)eof energy
(Rayner, 1991). Although Fairbaim includes infant experiences in his account of otject
base
his
he
did
(Bretherton.
1987).
theorising
on
actual
observations
not
infant
relations
In psychoanalytic object relations theory, self- and object-representations are

considered as multi -dimensional and complex, and the emergence of any particular
representational configuration reflects a synthesis of the imperatives of the external and
internal worlds (Sandler and Rosenblatt, 1962). Sandler and Rosenblatt (1962) draw a
distinction between the'inner world'and the 'representationalworld'stating that the inner
by
They
'representational
is
that
the
than
were guided
world'.
of
world a vaguer concept
from
developmental
a
a need to conceptuallse and categorise child clinical material
perspective.
In what is now considered a classic work, they state that the term
first,
two
organization
a
stable
psychic
concepts:
separate
implies
actually
representation
integrating
the
all
of
mental
images
and
collecting
map,
an
experiential
as
serves
which
the
the
between
dispositions
the
and
contents
second,
and,
others-,
and
self
and relational
dispositions
images
those
which
reside
within
and
of
characteristics
cognitive-affective
the
The
the
world
is
product
representational
of
construction
each personal experience.
functions.
of ego
Representationsare seento have an enduring existence as an organisation that is

that
gradually
with
about
come
or
images
of
many
impressions
a
range
Out
of
constructed
but
The
these
representations
contains
world
representational
experience.
and
maturation
in
body
from
the
the
with
interaction
its
own
child's
arising
also more, sensations
I
-)I
environment result in the fon-nation of a body representation.The representational world
is compared to a stage set within a theatre where the characters represent the child's
various objects as well as the child him or herself The self-representation is seen as the
organisation which represents the person the child has consciously and Unconsciously
perceived him or herself to be.
Another important aspect of Sandler and Rosenblatt's formulation of the

is
representational world that of the shape of self- or object-representations. By this it is
meant that the character and form of a particular representationor image is denoted at any
feels
For
the
example,
child
one moment.
who
angry at one moment, and is the subject
in
his
be
This
shows
a
change
of
shape
may
self-representation.
of attack at another,
identification
becomes
In
this
a modification of
model,
either conscious or unconscious.
the seIf- representation on the basis of another object representation as a model and
be
seen as organised changesin the self-representation.
enduring identifications would
Hans Loewald maintains yet revises Freud's structural model in a way which
(Kaywin,
1993).
Loewald
developmental
is
and relational perspectives
embraces
his
basic
because
discussion
the
for
regards
concept of
principles
this
of
one
important
is
intemalisation
him,
For
1989).
that
the
the
(Fogel.
activity
is
organising
internalisation
is
human
It
defines
the
a growth principle and
mind.
and constructs,
vcry essenceof, that
ioning
functi
basic
believes
He
de\,
inherent
the
the
tendency.
of
way
of
1
elopmental
an
is
to
the
generate
as
central aspect of its
representations
internalisation,
mind is
for
Intemalisation
Loewald
by
is a given
and revealed onk
its
instinctual activities.
by
internalisation
,
The
the
the
concept
of
represented
is
of
process
manif estations.
by
intemalisation
is
The
the
ego
characterised
process yet is also the
coherent ego.
product of that same process.
Loewald adds a unique perspective to ego development; that is, that id and reality
develop
the
are psychological constructions which, along with
ego,
over time. In
Loewald's conception of development, the early preoedipal period is the phase wlieri
id,
beginning
develop
to
of
ego
and
reality
are
out of an original
structures
psychic
Lid
develops,
Regarding
the
as a psychological construction which
primary narcissism.
Loewald (1978) says, "Understood as psychic phenomena or representatives, instincts
form
They
being
the most
the
early organising mother-infant interactions.
in
come into
life
In
human
level
their
totality,
and as mental
mentation and motivation.
of
primitive
different
levels
toward
of mentation and
more complex organisation of
progresses
from
distinguishable
the
between
the
them,
ego or
as
id
constitute
instincts
interplay
495).
(p.
superego"
What this means is that the instinctual drives themselves are not simply
biological-constitutional
interaction
in
forged
the
they
early
with
are
givens, rather,
drives
instinctual
defining
the
And,
1993).
than
(Kaywin,
as
mental
rather
environment
defines
did,
Loewald
Freud
them
biological
as
primitive
as
strivings
of
representations
is
fashion
before
in
become
develop
it
some
organised
and
must
which
mentation
id.
Instinctual
the
to
the
called
presenceof a psychological structure
appropriate to refer
develop
that
drives are ilow to be Understood as psYcliological
over t*ime
constructions
I
the
interaction
environment.
with
in
Loewald has preserved a drive theory while he
developmental
to
the
perspective (Kavwin.
revises
model infuse it with a relational and
1993; Fogel, 1989). Drives are psychological. representational - the products of
differentiation.
As stated Loewald believes that the differentiated psychic structure of id, ego and
develop
out of an undifferentiated state in which self and object
reality must subsequently
full
fused.
Freud's
Loewald
In
to
contrast
view of reality,
gives
psychological status
are
to reality. Reality also is understood as a psychological construction that develops over
time from early subjective forms to later more objective forms (Kay-win, 1993). He rejects
between
drives
(ego,
there
that
the view
organism) and
antagonism
is an inherent
by
of
the
out
mind
Drives,
created
are
(parents,
objects
and
ego
civilisation).
environment
full
internal
isation,
Full
human
in
integration is
relationships.
a context of
original unity
healthy
the
recovery of original unity as a goal of
a potentially realizable ideal development.
1.1.2
A Cognitive Approach
based
be
defined
to
Cognitive science may
effort
empirically
a
contemporary,
as
to
long,
those
efforts
explain
particularly
epistemological
questions
answci'
-standing
1:)
IlUman knowledge. Of the various features generally associated with cognitive scientific
human
bellef
there
the
that
to
talk
cognItIve
activItIes
Is
necessarv
It
is
cxPlorino
efforts
24
from
level
to
the
about mental representations and posit a
of analysis wholly separate
biological or neurological, on the one hand, and the sociological or cultural on the other
(Gardner, 1987).
Cognitive science has also made the deliberate decision to de-emphasize certain
factors which may be important for cognitive functioning but whose inclusion would
factors
"These
the
complicate
cognitive-scientific enterprise.
include the influence of
factors
affective
or emotions, the contribution of historical and cultural factors, and the
background
the
role of
context in which particular actions or thoughts occur (Gardner,
1987, p.6)11.
The cognitive scientist restshis discipline on the assumptionthat human cognitive

forms
in
be
described
terms
of symbols, schemas,images, ideas and other
activity must
level
legitimate
is
believed
It
to
and necessary posit a separate
of mental representations.
level
When
be
this
the'level
working at
of representation'.
of analysis which can called
in
deals
the representational entities mentioned, that is, symbols, rules,
the scientist
in
these
the
entities
are
joined,
representational
which
ways
explores
and
images,
Much
talk
theoretical
among
contemporary
another.
one
transformed or contrasted with
best
discussion
the
of
conceptualizing
mental
to
ways
of
a
amounts
scientists
cognitive
form
but
favour
Some
there
that
the
of
mental
a
single
is
view
representations.
in
least
believe
features
two
that
at
some
while
propositions,
one
usually
represcntation,
forms ofI mental representation- one more like a picture or image and the other closer to
One
least
(Johnson-Laird)
three
to
types
cognitive
scientist
at
prefers
posit
propositions.
of' mental representation: propositional representations: mental models which are
structural analoguesof the world: and images that are the perceptual correlates of models
(Gardner, 1987).
Based on observations of his own children, Jean Piaget's (1972) theory of

development centres on children's cognition and sensori-motor capacities. A child comes
to know or understanda rattle or other toy by acting on it, ph-sicallyand mentally. Rather
than passively taking in information,, Piaget believed that children have an active part in
the process of knowing; they actively select and interpret information in their
In
knowledge
this
environi-nent.
way, children's
of the world changesas their cognitive
development
is
develops.
Piaget,
For
the
essenceof cognitive
structural change
system
in
influences
its
the
thought
to
content
of
as well as
and
change
which gives meaning
information
knowledge
form
One
the
and
cast.
is
of
mental
in which
representations the most important changesin the structure of mental representationsand representational
processesis the advent of symbolic thought - the culminating achievement of cognitive
development in the infant. The ability to mentally represent one object or concept for
in
language
fantasy
for
basis
turn,
the
of
which,
the
acquisition
and
play
another provides
leads to greater cognitive development. Language and fantasy play also become the
begin
decipher
the
to
by
understand
and
can
which researchers and analysts
means
thoughts
and representations.
children's
of
contents
One particular
type of
mental
representation of
special interest to
devclopmentalists is scripts. These were conceived of b,,,,information-processing

I
introduced
Schank
best
formulation.
In
known
the
the
studying
adults.
ists
psNcholo(-)
notion of a script -a canonical set of events one can expect in an often encountered
"
setting such as a meal at a restaurant (Gardner, 1985). Such a structured framework
allows the 'understander' to deal efficiently with a variety of otherwise difficult to
assimilate texts. In other words, scripts are generalized, coherent mental representations
of a series of events that occur in a consistent temporal order in e'ery da\- life. TheNdescribe 'what's supposed to happen' in certain situations. According to Nelson (1978),
three features of scripts are noteworthy. First, both verbal and non-verbal representations
of self, objects and events are likely to form scripts which are established N'eryearly in
life. Second, the script approach seemsto correspond most closely to the way M \hich
children represent complex events of their day-to-day lives. That is, rather than a static
'event
to
picture', scripts consist of something closer an
schema' (e.g.,Mandler, 198-3).
Thirdly, because the approach is applicable to the social and emotional world of self,
is
their
particularly
integrating interactions and relationships, it
objects, events and
developmental
to
the
researchers.
appealing
Although both attempt to do so, neither cognitive theory, which focuses on

'rational' thought processesnor psychoanalytic theory, which tends to focus on 'irrational'
how
children's minds and
or unconscious thought processes,presenta complete picture of
be
However,
theory
develop.
the
may
actually
shortcomings of one
personalities
in
The
theory
by
the
cognitive
can
reality
of
external
role
the
other.
strengths of
remedied
fantasy
internal
theory
and
vice
versa.
psychoanalytic
the
in
of
role
inform or compliment
"'
1.2
ATTACHMENT
THEORY AND THE INTERNAL
WORKING
NIODEL
The ideas contained within the joint

work of John Bo-vlbvand MaryAMsxvortli
known as attachment theory are also central to this
project. Dra,ving on concepts froni
ethology,
cybernetics, information
processin(.1, developmental psN,

cholo(,,,
%, and
psychoanalysis John Bowlby forinulated the basic tenetsof the theory (Bretherton, 1995;
Eagle, 1995; Grossman, 1995).
Like other psychoanalysts, John Bowlby (1973) was greatly concerned

with the
between
relationship
outer reality and inner world. In many ways, his theories can be
seen as a bridge joining the complimentary aspects of psychoanalytic and cognitive
psychology points of view as examined above-, that is, that Bowlby's concept of'internal
be
link
between an individual's inner world and his or her
seenas a
working models'can
The
emphasis on unconscious representationsis most clearly expressed in
outer reality.
the centrality of the concept of internal working models (Eagle, 1995).
In reacting against central aspects of classical Freudian theory, both attachment

theory and contemporary psychoanalysis (e.g., object relations theory) end up sharing
basic assumptions (Eagle, 1995). For example, they both reject the idea that the infant's
is
latter's
infant's
to
hunuer
to
the
mother
secondary
the
attachment
role in gratification of
dri\ c and that the vicissitudes of the sexual drive are primary determinants of personality
dcvelopnient. The- share the views that infant-mother attachment is an autonomous
28
motivational system rather than one subservient to other instinctual driNes and that
is
personality
shaped by early attachment experiences.
Of particular significance to developmental researchersis Bo%%-Ib,,,

-'s suggestion
that the construction of internal working models of self and attachment figures is a
natural consequence of the human ability to make sense of the world through the
construction of mental representations. Through continual transactions with the world
increasingly
the
of persons and objects,
child constructs
complex internal working
1973,
(Bo%\-IbN,,
the
the
the
self
models of
world and of
significant personsin it, including
Bretherton, 1995b). The representationaldimension of attachment theory is basedon the
increasingly
become
that
the
a central
pattern of caring relationships
assumption
intemalisation
the
through
the
of early parent-child
child
cognitive-affective structure of
These
'internal working models' of caring experiences organize and
interactions.
development
the
throughout
relations
and
interpersonal
influence subsequentpersonality
life cycle (Blatt, 1995).
Bowlby suggestedthat within an individual's working model or representationof

figures
are of special
the world, working models of the self and of principal caregiving
significance.
history
based
the
develop
of
on
Infants
'internal working models'
the
They
the
'represent'
of
past
nature
only
their
not
caregivers.
primary
interactions with
future
forecasting
but
the
and
experience
of
they
permit
also
experience,
interactional
internal
future
That
working models represent
become the prototype of
relationships.
future
formation
for
become
the
the
of
experience
prototype
and
interactional
past
&Aber,
1992).
(Slade
them
the
transference
mechanisms
make
of
relationships
Ordinarily. internal working models have an adaptive function of providing an

adequaterepresentationof self, attachment fioures and environment. This means that the
t7
development of healthy attachment relationships is basedon the continual up-dating and
tine-tuning of the internal working models. The metaphor of the'intemal working model'
of the self and attachment figure underscores the dynamic and functional aspects ot
representations(Bretherton, Ridgeway & Cassidy, 1990). In infancy, the child's internal
figure
develop
in
working models of self and attachment
a complememary fashion and
Caregivers
initially centre around matemal availability.
who are able to read their infant's
increasingly
infants
by
'represented'
their
to
as available.
cues and respond sensitively are
In contrast, caretakers who tend to reject their infant's bids for attachment or who
insensitively
'unavailable'
to
them
to
as
provide
represented
are
increasingly
respond
being
formed,
At
the
these
time
the
are
as
representations
same
protection or security.
to
the
or
powerless
gain
self
as
powerful
of
representations
infant createscomplementary
felt
'
security' and eventually as worthy or unworthy of protection
a senseof security or
1992).
&
Aber,
(Slade
and care
Internal working models need not be fully accurate nor detailed to be useful but
be
important
it
functional
they
the
thev
that
fulfil
consistent
with
realitN,,
is
their
role
to
Under some circumstancesdefensive processesmay hinder
1987).
(Bretherton.
represent
Defensive
believed
internal
exclusion is
working models.
the adapti,,,e accommodation of
Clinical
intolerable
or
conflict.
to
pain
case
material
su(-,
mental
Lests
response
to occur in
figure
habitually
a
ridicules
child's
securityattachment
an
when
c,,
ari,,
that such conflict
secking,
lo%
feelillos
denies
disavo\
the
anxious.
an(-,
or
child*s
i-\
in(,
s or
behaviours or
defens'Nel-\
Under
figure.
these
a
child
\\ould
circumstances
towards the attachillent
exclude from awareness the working model of the 'bad' parent and retain conscious
access to the model of the *good' loving, parent. As the Internal workIng model of the
unconditionally loving parent does not correspond to realItYsuch an idealized model is
maladaptive. Although the defensive processes might bring relief from psychic pain,
inadequate working models will interfere with effective coping and with optimal
development (Bretherton, 1987; Brether-ton,Ridgeway & CassidY. 1990).
The biological 'set goal' of the infant attachment behavioural system was
by
Bowlby as physical proximity to the parent in order to ensure
originally conceived
Later,
the psychological set goal of 'felt security' was added to
and
survival.
protection
behaviours
beyond
Thus
to
the
attachment
infancy.
concept relevant
when the
make
familiarity
by
is
feeling
to
mother, or
with the
safe and secure, virtue of proximity
infant
behavioural
the
system or the need to signal mother to comfort
attachment
environment,
feels
is
By
deactivated.
the
child
in need of comfort,
contrast when
or provide safety
because of mother's distance from him or her or the perception of danger from the
be
When
level
felt
a critical
activated.
of
environment, the attachment system will
(Slade
1992).
Bw,
Iby
Aber,
deactivated
been
has
the
and
,
systern is
achieved,
security
base
to
their
children's
exploration
of
inner.
as
pertaining
secure
of
the
role
parent
saw
1995c).
(Bretherton,
not just their outer world
Research indicates that insecure
(anxious
with
relatedness
ambi%alent
preoccupation
a
in
either
Lllts
attachment I'C-,
attachment)
distorted
(avoidant
emphasis
on
separation
insecure
oi- in an exaggeratedand
1995).
(Blatt,
attachment)
The next section will describe

for
I
assessim-,
a .vel -established empirical method
quality of attachment in infants that was used in the first phase of this studv (to be
described in Chapter 2).
1.2.1
Assessment of attachment in infants
TheStrange Situation
Originally employed in 1970, the Ainsworth Strange Situation (Ainsworth,

Blehar, Waters & Wall, 1978) has been established as a reliable and valid empirical
instrument with which to assess the quality of infant-mother (and infant-father)
from
drawn
Based
conclusions
prolonged observations of mothers and
on
attachments.
differences
in
home,
Ainsworth
that
the
reasoned
in maternal sensitivity to infant
infants
in
differences
life
first
during
to
the
ought result
in the quality of the motheryear of
cues
laboratory
based
This
assessmentinvolves
infant attachment relationship at one year.
Focus
the
two
parent.
reunions with
t,vo separationsof parent and child and
is upon the
infant's behaviour, particularly during the reunions, vvbere indi,,,idual differences are
in
to
this
the
cope
with
stressful situation
terms
employed
strategies
of
incaSUred
(A insworth et al, 1978).
follows
brief
the
A
classifications of infant attachment
and is
overview of
Kaplan
&
Main,
1996S'lade
&
Aber.
(George.
necessarik condensed and simplified
I 992).
32
Secio-elyA ttached Infants (B)
Most infants assessedin the StrangeSituation demonstratea pattern of attaclu-nent
behaviours
that Ainsworth believed to be eNidence of a 'secure'
and exploratory
attachment relationship (about 65% of white, middle class American and British infants
infant
On
this
the
the
the
may
exhibit
pattern).
entering
stranger
playroom and seeing
initial
be
by
However,
(this
to
wariness.
considered
checking with mother
is
show
displaying securebase behaviour) these infants derive sufficient security to explore their
infant
departure,
When
leaves
her
the
the
the
parent
secure
registers
room,
environments.
for
diminished
by
looking
her,
by
perhaps
quality of exploration and
perhaps crying or
infant
The
will not use the stranger as a substitute play partner and will not carry on
play.
distracted
by
be
before
the stranger.
the
somewhat comforted or
infant may
although
as
When the parent returns the infant may signal or greet the mother actively or allow
derive
from
They
be
to
themselves
clearly
comfort and security
picked up and comforted.
infants
Parents
to
to
securely
of
attached
mother's return and are able return exploring.
have been found to be sensitively and contingently responsive to their infants' cues.
41l.vio IIs-,.I voidant Infants (A)

The largest group of children that fall into the insecure category are described as
having an anxious-avoidant attachment pattern. Prior to separati
infants
avoiIdant
are
from
interested
distinguish
in
difficult
they
too
to
secure
children
as
are
somewhat
do
However,
they
the
not appear to need a warm-up period
)
iiew em-ironment.
I
cxplormLg
do
to oNercome %\ariness. not check in much with their parent through proximity seeking
When
leaves
do
the
the
the room avoidant children
stranger.
mother
and not seem war-of
do not seem to protest and there appearslittle change in the level of their play. They may
find it relati,,,ely easy to substitute the stranger for mother as playmate. It Is upon the
infants.
from
distinguished
As
the
secure
that
easily
avoidant infants are most
reunion
her.
There
infants
either conspicuously avoids or ignores
mother approaches, avoidant
distress
little
and no anger and the response appears
or no proximity seeking, no
is
is
focused
The
on the toys or environment throughout the
child's attention
unemotional.
found
been
be
have
Parents
to
predictabl,N,unresponsi\*e
of avoidant children
procedure.
free
for
bids
20%
Approximately
to
comfort and are controlling in
play.
and rejecting
infants
insecure-avoidant.
British
American
and
are classified
of white middle-class
Anxious-Resistant Infants (C)

A small number of children show a pattern of attachment behaviours that is more
They
are called anxious-resistant or
or
maladaptive.
easily interpreted as non-optimal
be
distressed
These
children may
wary or
even prior to
anxious-ambivalent children.
little.
They
they
seempreoccupied with the parent throughout
explore very
separationand
the procedure and may seem angry or passive. Upon reunion with the parent, the child
tails to settle and take comfort and continues to focus on the parent usually crying in an
fail
Parents
They
distressed
to
to
return exploration or play.
manner.
of resistant
angry or
bids
for
to
comfort and are
children are unpredictably responsive and inconsistent
in
free
play.
unresponsive or unavailable
Roughly, 10-14% of white. middle-class
American and British infants are judged to be avoidantly attached to their mothers.
Di. sorganisc'I
d
(D)
CILS()rieflt(!
I-his category is the most recently identified. Approximately 12% of infants in
34
fall
liowever.
British
American
this
group,
into
samples
and
white. middle-class
These
high-risk
(Cassidy,
1994).
into
fall
infants
this
samples
group in
substantially more
in
disoriented
behaviours
disorganised
the
that
parent's
presence
and/or
show
infant's
in
face
being
behavioural
lapse
the
the
attachment system
strategy
of
of a
suggest a
triggered. For example, the infant may freeze with a trance-like expression or they niay
floor.
fall
huddled
to
the
the
prone and
on
parent's return only
rise at
Mary Ainsworth's work tended to emphasise the behaviour of the mother-child

described
however,
different
the
that
the
assumption
as
above,
underlying
is
couple,
from
Situation
derive
different
histories
Strange
the
relationship
patterns of attachment in
different
underlying representations of the relationship (Slade and Aber,
and reflect
1992).
Attachment theory can be thought of as consisting of two parallel but interrelated

dimensions: 1) an interactional or behavioural dimension, concerned with the
development of and maintenance of particular patterns of behaviour that are established
interactions
2)
dimension
based
a
or
and
representational
cognitive
in infant-parent
on
development
This
dimension
theories.
and psychoanalytic
second
cognitive
of attachment
theory is concerned with how intimate personal interactions are established or
internal
in
how
the
as
of
mind
these
working
models
caring
experiences
internalised
and
become
increasingly
a central cognitive-affective structure of the child through
schema
the internalisation of early parent-child interactions (Blatt, 1995).
1.2.2
A move to assessing the representational level
As stated above, the representational dimension of attachment theory is based on

the assumption that the pattern of caring relationships increasingly become a central
cognitive-affective structure of the child through the internalization of early parent-child
insecure-avoidant
The
of
secure,
patterns
and insecure-resistant attachment
interactions.
fundamental
influence
behaviour
in
that
cognitive affective schemas
not only
establish
forrn
basis
for
in
but
behaviour
the
also
nonnal and abnonnal
adolescenceand
childhood,
adulthood (Blatt, 1995).
The development of methods to assessattachment patterns through narrative

focus
has
the
of attachment researchers to an exploration of mental
expanded
reports
findings
These
have
models
of
attachment.
structures or internal working
redirected the
focus of attachment theorists to the structure of the representationalworld (Diamond and
Blatt, 1994).
It is the case that parental behaviour, however subtle, mediates the relation
betweena parent's stateof mind with respectto their own attachmenthistory and their
behaviour
Strange
Situation
(George,
Kaplan & Main, 1996).
toward
that
parent
infant's
Fona, N, Stcele and Steele (1991) have demonstrated that, within the sample referred to
-,
before
the
thesis.
this
of
the birth of the child
parents'
assessment
attachment
status
in
father
to
to
classification
attachment
mother
at
one year and
predicted
at eighteenmonths
&
Fonagy,
Steelc
1996).
(StcLlc.
36
The Adult A tiachment Interview
The primary method for assessingadults' internal working models of attachment

is the Adult Attachment Interview (George, Kaplan & Main, 1996). This interview is an
hour long, semi-structured interview focusing upon the description and evaluation of
The
Adult
Attachment
and
attachmentrelated experiences.
early attachment relationships
Interview has been shown to be stable acrosstime, unrelated to intelligence, unrelated to
both short-term or lang-term memory and unrelated to social desirability and not
Subjects
interviewer
(George,
Kaplan
&
Main,
1996).
to
are asked
effects
attributable
directly about childhood experiences of rejection, being ill, hurt and upset as well as
focuses
The
loss
the
assessmentof
interview
separation and abuse.
about experiencesof
description
the
participant's
and
on the overall coherency and consistency observable in
evaluation of childhood experiences,and upon the participant's ability to collaborate with
the interviewer by giving complete but succinct answers,remaining on the topic enquired
The
interviewer
by
the
aim of the
and speaking in a clear non-confused manner.
about
but
the
the
to
to elicit
experiences
of
subject's
childhood
actual
interview is not appraise
their representations of the meaning of early experiences through their reconstructions
interview
Once
the
to
the
text
the
telling.
interview
scores are assigned
is considered
in
delineates
light
the
several possible 'states of
of a classification system which
again in
four
There
to
attachment.
are
categories of 'states of mind with
mind' \vith respect
best
fit
that
to
the
to
to one Is primary
are
considered
attachment'
provide
respect
attachment relationships and attach-ment-relatedexperiences and are widev in use. They
are
known
as
the
secure-autonomous,
dismissing.
preoccupied
and
37
brief
&
Main,
Kaplan
1996).
(George.
A
overview of
unresolved/disorganised patterns
the classifications follow.
Secure-aulonomous (F)
The interview of the secure-autonomous adult is characterised by coherent
discourse. These adults are valuing of attachment but seem objective regarding Lmy
descriptions
Their
and evaluations of attachment-related
particular event or relationship.
favourable
is
the
experiences are
or unfavourable.
experiences consistent whether
Dismissing (D)
These interviews are characterised by incoherent discourse. The subjects are
dismissing of attachment related experiences and relationships. They tend to normalise
history
their
unsupported or
people and experienceswith generalised representations of
The
be
by
transcripts
tend
to
autobiographical episodes recounted.
actually contradicted
excessivel,,,,short.
Preoccupied (E)
These interviews are also characterised by incoherent discourse, however, these
by
past attachment relationships and experiences. The
subojectsare preoccupied with or
fearful.
Their
angry.
passive
or
even
appear
sentences are often long,
speakers often
long.
the
transcripts
or
and
vague
entangled
excessively
grammatically
Unresolved disorgamsed (U/d)
38
The unresolved interview is identified mostly during discussions ofloss or abuse
discourse.
lapse
the
their
monitoring of reasoning or
when
individual shows a striking
in
For example, the individual may briefly indicate a dead person is believed still alive or
killed
The
thought.
that
they
subject
childhood
may momentarily indicate
someonewith a
fit
may otherwise
into one of the above categories.
1.2.3
Recent perspectives on attachment theory
In the beginnings of attachment research, researchers primarily focused their

investigations on the elucidation of attachment processes in infancy and on the
developmental sequelaeof secure and insecure infant- care-giver attachments.
More recently, researchershave conceptualised attachment as remaining critical

to the child's continual adaptation even though it changes in organization as the child
develops throughout the preschool period; this is in keeping with Bowlby's perspective
(Cicchetti, Cummings, Greenberg & Marvin, 1990). The need to understand the
developmental transformations of attachment processes in the preschool years has
become increasingly critical. The move to the representational level described above and
the concept of
figures
internal working models of attachment
and of oneself has
implication
The
be
these
to
to
investigate
questions.
ability
that
our
seems
cnhanced
for
be
the
to
should
most
part,
not
continue
conceived of as an autonomous
attachment.
39
its
but
that
area of study
effect on later developmental tasks should also be addressed
(Cicchetti et al, 1990). Accordingly, theorists and researchers have recently begun to
different
from
described
the
types
examine
of attachment relationships
slightly
above
different perspectives. Two of these are relevant to this thesis and are described below.
ommunication
erspective
The communication perspective of attachment rests on the assumption that

initially
behaviouralthrough
to
their
transmit
own patterns of relating
children,
parents
later
dialogue
but
future
through
also
verbal
about past,
and
affective interaction patterns,
hypothetical experiences (Bretherton, 1995c).
Open and effective communication has been found to have links between parents
infancy,
toddlerhood and young children
and children and secure attachment in
(Bretherton, 1987). (See Chapter Two for more on this) For example, Main and
being
in
describe
(1985)
the
at ease exploration of
secure six-year-old as
colleagues
teelings and potentialities. Secure parents, like their children, were characterised by the
ease with which positive and negative aspects of attachment experience were
communicated and integrated.
In order to clarifv this underlying assumption about the theoretical connections

bet\\een emotionally open communication and adequate, adaptable internal workino
following
illustration.
Bretherton
(1987)
the
gives
models
If an attachment figure
40
infant
learns a senseof trust that
the
to
appropriately responds security- seeking signals,
balance
behaviour.
If
the attachment
attachment
results in an optimal
of
and exploratory
figure does not respond appropriately to the infant"s signals, the infant %vIII fcel
dissatisfied and misunderstood or disavowed. This, in turn, has consequencesfor the
internal working model the infant will construct of self and the attachment figure.
Mutually satisfying communication is impossible without shared working models. In
dyads,
attachment signals are mutually responded to and understood and so the
secure
fine-tuning
Such
to
are
adequate
and
also
open
and
models
up-dating.
internal working
patterns of communication may retain their qualitative stability, even though the child's
become
(Bretherton,
1995c).
more complex
communication skills
AtfL,ct Regulation
Another of these more recent ways of looking at attachment patterns, which is
described
has
the
to
above, is that
communication perspective
overlaps with
related and
been
has
proposed that emotion regulation and quality of
of emotion regulation and it
This
linked
1994).
(Cassidy,
recent conceptualising speculates
Littachment are closely
from
be
differentiated
basis
insecure
the
each
other on
that
and secure organisations can
definition
Central
1993).
(Slade,
to
the
of emotion
of affect regulation strategies
heightening
involves
both
the
that
and
of emotions, the
suppression
it
re,()Lilationis
factors
both
involves
(temperament)
and
and
intrinsic
extrinsic
attention,
of
reoulation
t,
(particularIv the child's relationship with
(Thompson,
1994).
parents)
Emotion
be
by
the attachment relationships through the child's
to
thought
influenced
re,pilation is
behaviour.
For
is
the
of
models
parents
or
\vorkingl
example,
secure child
cxpectations
his
her
develop
that
to
an expectation
or
emotional signals will be responded to.
thought
41
Because the parent is sensitive to the child's signals, both negative and positive affects
freely
be
during
the
times of'
expressedand experienced as useful in alerting
parent
will
distress. Flexible emotion expression could be seen as part of a strategy aimed at
freedom
1994).
The
(Cassidy,
to
the
explore
while
assuring
safety
pattern
allowing
infant
of openness to a range of emotions seen in secure infants is also evident in the Adult
Attachment Interviews of secure-autonomous parents regardless of the quality of the
being
discussed
(Cassidy,
1994,
George,
Kaplan
&
Main,
actual childhood experiences
1996).
One of the hallmarks of insecurity appears to be an individual's inability to

integrate negative affects, memories and
thoughts into their representations of
themselvesor their parents (Slade, 1993). In fact, difficulties regulating and modulating
insecure
be
fundamental
there
to
to
the
organisations and
appears
a
negative affects are
distinction between minimising and maximising strategies (Cassidy, 1994). It has been
proposed that when activation of the attachment system consistently results in rejection,
develop
the
to
these
attachment
attention
relationship;
of
minimising
a
strategy
infants
described
(Main
&
Solomon,
1986).
as
above
as
avoidant
infantsaremostoften classified
In order to minimise the attachment relationship a strategy of minimising negative
distress
be
it
could
useful as avoids rejection of
emotions such as anger, sadnessand
infant
behaviour
the
sufficient proximity to the parent
\vhile also permitting
attacl-ii-nent
tIor safetv (Cassidy, 1994). Positive affect may also be minimised as it engages the
for
tigure
open interaction.
and signals a readiness
attachment
Int'ants classified as insecure-resistant or ambivalent displaY an-gyr.

y. resistant
42
behaviour toward the parent in the Strange Situation.
What attachment related
lead
to a strategy of heightening emotions ? An infant who has
experiences might
experienced minimally or inconsistently available parents needs to develop a method of
heightening the importance of the attachment relationship. Heightened negative
be
emotionality can
viewed as a component of the child's strategy to gain the parent's
be
attention and may
chronic becausethe child recognisesthat to relax and allo,.v him or
herself to be soothed by the parent is to run the risk of losing them (Cassidy, 1994).
Both insecure strategies described above seem to be manifestations of a failure

to integrate negative affect, yet they function in different ways. Dismissing or avoidant
feelings;
preoccupied or resistant individuals exaggerate
individuals minimise negative
them (Slade, 1993). The presumption is that these strategiesdevelop becauseof the need
to preserve the primary attachment relationship. That is, infants whose parents are
uncomfortable with negative affect will avoid it and those whose parents are overly
because
few
However,
there
to
are so
negative affect will exaggerate it.
responsive
is
little
data
demonstrate
insecure-resistant
there
to
empirical
a
infants classified as
heightening of negative affect in these infants. There is evidence of minimisation of
i
dant
insecure
(Cassidy,
1994).
the
pattern
nezativcaffect in
-avo
43
1.3
CONCLUSIONS
This chapter has reviewed the literature concerning the main ideas relevant to this
thesis such as mental representations, internal isation, attachment theory and internal
is
develop
What
to
orld
needed
is
measuresof assessingthe interrial, ,,,,
working models.
from,
derived
to
the
the
or are
attachment system which are
young child with regard
of
developmental
tasks and
theory,
the
yet reflect
new skills,
consistent with, attachment
Greenberg
&
Cummings,
(Cicchetti,
the
growing child
physical-social environment of
Marvin, 1990). This issue will be examined in greater detail the next chapter.
44
CHAPTER TWO
ASSESSING THE INTERNAL
WORLD OF THE YOUNG CHILD
45
2.1
INTRODUCTION
The previous chapter reviewed literature conceming
various theoretical
internalisation.
internal
working models
perspectivesregarding mental representations,
For
Chapter
One
theory
to
this
thesis.
pertinent
example,
revie,ved
and attachment
ideas
interrialisation
theories
and
more contemporary
regarding
classic psychoanalytic
in
detail,
It
the
also examined,
some
attachment theory's
representational world.
and
internal
the
working model and empirical methods with which to assessthem.
notion of
The current chapter will continue the theme of exploring the internal world by looking
at past and present empirical attempts at assessingthe young child's internal world.
brief history of investigations using the technique of doll play, mainly by cognitive
followed
by
This
the
the
a
review
of
is
use of and
chapter.
opens
psychologists,
by
The
the
to
child psychoanalysts.
chapter
inner world
access
understanding of play
literature
the
then
concerning assessingthe internal working models
empirical
review
will
in
be
instrument
It
this
to
the
employed
Investigation.
of young children and introduce
known
MacArthur
Story
Stem
Battery
developed
focus
the
II
technique
as
on a recently
wi
(MSSB) (Robinson, Mantz-Simmons, Macfie & The MacArthur Narrative Working
Group, 1992). This chapter will also introduce the design of the current study and the
MSSB
to
the
to
children about whom much early
apply
allows
it
opportunity
unique
history is known.
46
2.2
DOLL PLAY TECHNIQUES
Doll play has been used both by researchersand clinicians such as child analysts
internal
for
therapists
the
as
a
means
of
gaining
and play
understanding of
child's
world
decades.The work of R.R. and Pauline S. Searsat the Iowa Child ResearchStation in the
doll
in
940's
development
I
the
(Levin
as
a
utilised
play
method
study
of
personality
mid& Wardwell, (1962). There are numerous variations on the theme of doll pla-y, but
I
family
is
dolls
the
presented with a set of
young child
essentially
- and a
- usually a
house
dolls
in
to
told
the
the
to
manipulate
as
a
and
are
are operate - such
setting which
dolls while the child tells a story about them. Becauseof the theoretical disposition of the
from
derived
behaviour
frequent
the
were
variables measured
most
carly investigators,
theory and were indices of acquired drives in children. Hence, more than any other
behaviour fantasy aggression has been measured by this technique (Levin & Wardwell,
1962).
2.3
CHILD PSYCHOANALYSIS
Child psychoanalystshave long viewed play as a window into the inner life of the
fantasy
that
to
thought
configurations
are
represent
are
child and play activities
Cicchetti,
(Marans.
Mayes,
in
to
uppermost the child's mind and accessible expression
DAL Marans & Cohen, 1991). In psychoanalytic formulations, children's play serves
fulfilment,
function
assimilation. reduction of anxiety and also
the
of i-naster-\,\\ish
have
frustrating
family
that
a
or anxiety-raising, meamng
events
commumcates current
47
tor the child (Freud, A. 1946; Marans, Mayes & Colonna, A., 1993. Solnit, 1987). The
focuses
on the specific themes in the play, while simultaneously attending to other
analyst
domains such as accompanying affects and changesor disruptions in play. It is out of the
from
the
observations
synthesis of
many domains that hypothesesare generatedabout the
developmental
child's
status and the dominant concems and intra psychic conflicts
(Marans et al, 1991).
Thus, observations of play activities emerging in the consulting room have a

diagnosis
However,
the
treatment
and
psychotherapeutic
of young children.
central role in
in
diagnostic
have
the
therapeutic
there
the
of
play
and
process
central role
in view of
been surprisingly few systematic investigations of the specific play themes, modes of
lead
to clinical inferences and
that
the
methods of observation
presentation, and
hypotheses. In a study by Marans et al (1991) attempts were made to addressthis lack
for
by
developing
assessingwhat
a systematic and reliable method
of empirical evidence
emerges in the consulting room with children.
In particular, the study developed a technique for tracking and marking the
in
These
during
themes
an analytic session.
a child's play
appearance of specific
development
by
theory
and included
and child
psychoanalytic
concepts were informed
fighting
damage,
bodily
and attacking and
rejection, reconciliation,
such items as
fourfirst
fixing.
The
tested
to sixon a non-clinical group of
system was
cleaning and
After
refining the coding system and improving observer reliability, results of
year-olds.
tive children are reported. This sample is quite small and so results must be treated with
i
ity,
I
however,
doubt
their
to
the
sabi
general
as
authors raise some relevant points
some
48
for the present study. Regarding the themes the five children used, they clearly alternated
between themes related to setting the scene and those relating to dramatisation of the
fantasy. More importantly, the authors found that analytically informed clinicians can
agree with moderate reliability about the predominant themes presentedby a child during
instruction,
that
with revision and
play interviews and
agreement among observers using
this technique generally improves. This is relevant to the present study as the main tool
be
has
little
data
inter-rater
to
used
published
regarding
reliability and
of investigation
the extent to which themes can be operationalised well enough for rater reliability is one
of the questions this thesis will address.
2.4
ASSESSING INTERNAL
WORKING MODELS OF YOUNG CHILDREN
As referred to in the previous chapter, a central focus of attachment researchover

from
based
better
been
has
last
transition
the
to
sensorimotor
ten years
the
understand
during
to
cognitively
infancy
relationships
attachment
representations of
based
&
Sroufe,
1997).
This
has
Carlson
(Fury,
representationsof attachment in older children
been attempted in various ways. For example, Kaplan and Main (1985) were the first to
In
drawings
might capture representationsof attachment. a recent
suggestthat children's
8-9
high-risk
year-olds,
study of
children's representational models of self and
investigated
family
drawings
figures
(Fury
1997).
The
were
in
et
al,
results
attachment
draxvincys
for
into
the
their
tapping
Supported
use
of
as
a
measure
gencrally
ofthe study
t
childrcn's repi-csciitational models of attachment.
49
2.4.1
Representation and narrative
Slade (1997), referring to clinical work, has suggested that for manv children
involves
by
therapy
the child and therapist to co-construct out of the
of
attempts
much
child's conflictual and often chaotic experiences, emotionally coherent narratives. Once
formed.
she argues, children are more capable of regulating their
such narratives are
behaviour.
Narratives
intra
their
as
as
well
emotions
are not only viewed as
psycliic
by
in
isolation
but
fundamentally
children
as
structures constructed
interpersonal,
in
developing
has
been
It
the
context of significant relationships.
suggested
emerging and
that narrative capacities introduce a new level of self and emotional regulation (Stem,
1985; Wolf, 1990).
The link between the internal world,
becomes
and
representation
narration
developmental
for
both
researchers when we
psychoanalytic and
particularly salient
in
it
functions
be
the
the
that
the
way
which
may
mind and
organisation of
consider
hold
for
discourse,
forms
that
this
true
and
may
of narrative
revealed through particular
ideas
One
Chapter
the
theory
of
attachment
and
introduced
as
adults.
well
as
children
It
'internal
the
the
the
that
concept
of
internal working
model'.
working
of
specifically
Adult
the
theoretical
the
narrative
research
such
as
that
used
in
underpinning
is
model
Attachment Interview described in the previous chapter. Clinical psychoanalytic data are
basedon autobiographical narrations told by the patient to the analyst. Similarly, a large
based
on stories or narratives that children
number ofexperimental studies of children are
families
during
their
a structured pla-v situation. And, coming
tell about themselves and
from
begin
form
these
to
t1ill circle, it is
childhood which
naiTatives
an autobiographical
BIBL
LONDON
UNIV
50
into
life
history a patient first presents to the
that
the
ultimately evolves
accounting
analyst.
According to Daniel Stem (1985), the making of a narrative is not the sameas any
kind
The
thinking
talking.
or
of
making of a narrative involves thinking in terms
other
intentions
and goals that unfold in some causal
of persons who act as agents with
beginning,
building
Stem
to
this
a
a
with
middle
and
an
end.
as
a
Lip of
sequence
refers
it
important
developmental
domain
'narrative
that
and
suggests
constitutes
an
in
self
a
fon-n.
The
the
translation
that
of personal experience into narrative
it involves
children in
idea that personal experience might take on the character of a narrative construction was
his
Screen
(1899),
Memories
Freud
In
in
Freud's
essay
earliest works.
one of
implied
from
have
our childhood; memories relating to our
any memories
questions whether we
be
all we possess.
childhood may
Bretherton, Ridgeway and Cassidy (1990) draw our attention to the idea that play
internal
for
be
language
working models
assessingchildren's
useful vehicles
could
and
They
but
in
that,
the
caution.
note
as
urge
relationship
attachment
of self and other
has
language
have
(1985)
Stern
(1980)
Bowlby
a curious relationship
pointed out,
and
how
be
to
A
guidance
on
given verbal
verbal child can
to internal working models.
interpret specific interpersonal events and thus acquire internal working models
the
This
experience,
child's
non-verbal
clarify
may
information
secondhand
'IcariOLISJV.
bUt it may also be at odds with it.
Bowlbv's
ideas about the importance of parent-child communication
in
in
in
(particularly
development
trends
the
are
reflected
recent
assessment
attachment
in
in
Chapter
1),
to
especiallly the use of children's
communication perspective referred
(Oppenheim
&
Waters,
1995).
Buchsbaum
to
attachment
status
assess
narratives
and
Emde (1990) feel that an important aspect of how the child makes sense of the world
important
A
telling
the
of
a
story
about
process
oneself
involves
and others. most
aspect
is
this
of
process that the child shares experiences with others; In this waY, the child's
influence
lines.
developments
Both
to
the
or'co-construct'the
caregivers come
story
in
language and the caregivers' increasing influence through narrative 'co-constructions' of
important
for
development.
are
many
aspects
of
a
experiences
child's
As stated, caution has been advised when using language as a vehicle for
assessing children's internal working models of self and other in the attachment
if
language
Emde
(1990)
Buchsbaum
the
onset
of
narrative
wonder
and
relationship.
introduces a disjuncture in the child's experience. Winnicott (1965) thought that the 'false
language.
by
from
the
use of
socialization experiencesand was made possible
self arose
Bowlby, and later Stern (1985), extended this idea by making a formulation that ties a
language
According
domain
to
this
to
onset.
view, the
self-experience
of
separate
is
distortion
distant
from
is
domain
to
and more
a core
susceptible persuasive
narrative
direct
Once
from
built
emotional
experience.
and
more
prior sensorimotor
up
self
language occurs, it is possible that children present more socially acceptable responses
feelings.
truer
their
to moral and affective situations and mask
52
In the first study designed to assessinternal working models of children frorn a
from
forty
Cassidy
Main,
Kaplan
(1985),
elicited responses
and
narrative perspective,
father
in
drawings
(the
to
separation related situations
of mother,
and child
six-year-olds
Separation Anxiety Task).
The six-year-olds who had been classified as securel-
infancy
gave embellished, coherent and open responsesto the pictures and
attached in
tended to volunteer information regarding their own separation experiences. Subjects
but
in
infancy
described
the
could not
children as sad
with mother
classified as avoidant
insecuredo
Children
to
the
the
situation.
classified as
cope with
children might
say what
disorganised were often silent or gave bizarre responses.
Before describing studies using the play narrative technique to assess the
in
the
the
the
coas
partners
of
adults
role
child,
young
of
representational world
Children's
be
to
talk
abilities
considered.
must
construction of children's narratives
in
issues
their
the
of
conversations
with
context
emerges
about emotional and personal
(Oppenheim,
Emde
both
child and parent
parentsand represent the joint contributions of
& Wambolt,
1996). Recent empirical work has demonstrated how parents help their
different
how
parental styles influence
children structure or construct play narratives and
Emde,
Wambolt,
1996;
(Oppenheim,
the
the nature and quality of
play narratives
Oppenheim, Nir, Warren, & Emde, 1997).
Oppenheim et al (1995), in a recent study of parent-child co-construction of

between
level
inter
the
associations
showed
of
subjectivity
affect and conflict narratives.
dyadic
the
constructions
and
coherence and number of prosocial
in
affcct
and shared
53'
themes in the children's independent narratives. In other words, the ability of children
in
facilitated
both
the
the child's
to
and
affect
meaning
co-construction
and parents share
independentnarrative construction and there was evidence of separatecontributions from
in
father
Their
the
child's
narratives.
and
work emphasising co-construction
mother
in
development
internal
the
of
processes
working models highlights the lack of
how
theory-based
studies of
attachment
parents communicate with their children about
(Bretherton,
1995c)
experiences
affective
2.4.2
The MacArthur
Story Stem Technique
The studies reviewed above document interesting connections between sensitive

by
attachment partners within parent-child
and emotionally open communication
to
talk
about attachment
capacity
separate
each
partner's
and
relationships
attachment
Oppenheim
&
Waters,
1995).
1995c;
(Bretherton,
issues openly and coherently
The MacArthur Story Stem Battery ( MSSB; Robinson, J., Mantz-Simmons,

Macfic. J. & The MacArthur Narrative Working Group, 1992) is a relatively new
the
representational world of children,
into
technique
a
window
provides
which
research
The
MSSB
family
themselves.
their
of
and
views
relationships
their understanding of
be
that
they
standardised
so
way
may
coded
c1licits children's play narratives in
&
Emde,
Kelsay
1998).
Klitzinu),
(von
systematically
54
This method has been used in a number of studies in addition to the codescribed
but
few
have
above,
construction studies
attempted quantitative analysesof the
data. For example, one qualitative study compared different stories between a sample of
maltreated and non-maltreated children (Buchsbaum, Toth, Clyman & Cicchetti, &
Emde, 1992).
Story stem techniques were first used to examine children's moral development
(Buchsbaum and Emde, 1990; Woolgar, 1996). Since morality is influenced by
interactions with significant others and involves dealing with the roles of others in the
midst of conflict, the story stems were constructed in two areas. One group of stems
for
development
for
themes
the
the
narrative
in
area
of
moral
probed
and
other group
themes in the area of family relationships. Buchsbaum and Erride (1990) found that by
36 months of age children were clearly able to produce narrative representations of
individual
'rich
There
themes
themes.
common
as well as
was evidence of
emotional
behaviour,
domains
the
adherenceto rules, reciprocity,
of empathy, prosocial
variation in
indicated
findings
family
The
the potential of this technique
relationships'.
and aspectsof
tIor obtaining data about children's internal and external experiences.
Bretherton, Ridgeway and Cassidy (1990) reported their findings of a study

assessing the internal working models of attachment relationship in three-year-old
The
designed
individual
technique.
the
to
story
stem
stories
elicit
were
children using
differences in the children's enactment of a variety of attachment-related issues. Thev
devised ,.i systern for classit'ying the story completions as Indicative of secure or insecure
attachment and compared this to a variety of other assessments.Detailed verbal and

behavioural transcriptions were made of the video-taped story completions and the
transcripts were subsequently analysed in two ways. First, the content was analysed in
order to examine the children's ability to understand the story issues and to create a
Second,
each transcript was considered as a whole in order to classify the
resolution.
children's story presentations as reflective of secure or insecure attachment patterns.
The authors established separatecriteria for security for each story. For example,
dealt
'monster'
fear
the
the
the
story responseswere coded secureif
parents
with
child's
in
for
Two
the
the
the
types of criteria were
or
child
approached
parent
comfort.
of
monster
issue
insecure.
If
for
the
the
the
as
child
avoided
main
of the
responses
categorizing
used
insecure
Avoidance
score
was
given.
of the
story or gave odd or incoherent responsesan
insecurity
because
the
avoidant
authors
of
as
indicative
regarded
was
story issue
issues.
Odd
defensiveness
behaviour
to
to
and
attachment
with respect
attributed such
disorganized responseswere regarded as indicative of a different type of insecurity. It
for
ambivalent
is interesting to note that no consistent pattern of responsewas suggested
children.
The authors concluded that coherent and emotionally open responding to
in
three-year-olds
securitv versus
predicting
successful
were
stems
story
attachment
the
at
same stage.
procedure
separation-reunion
a
in
insecuritv
The above findings indicated the potential of the story stem technique for
different
data
the
this
nternal
world
and
aspects
of
young
child's
it
is
about
obtainino
be
employed in this study.
tcchnique which will
56
2.1.5
MacArthur
Narrative Coding System
How one codes children's play narratives depends on the questions researchers
in
Klitzing,
Kelsay
&
interested
(von
Emde,
1998).
For
example, the rating approach
are
described by Marans et al (1991) above approximated the clinical process which makes
in
the
that
concerns
are
uppermost the child's mind at the time of the
inferences about
is
play interview and one which attempts to synthesisemarking the appearanceof specific
themes which are defined by a combination of discrete behaviours, narrative content and
Warren
developed
(1996)
As
affects.
mentioned
above,
and
colleagues
accompanying
focused
on children's emotionality and self-representations.
a coding scheme which
Two broadly compatible, but distinct, approaches to conceptualizing internal

date:
been
have
infancy
to
taken
representational approachesand
working models after
Representational
I
approaches analyse children's verbal
regu atory- style approaches.
drawings
themselves
of
and
children's
to
and
stems
story
response attachment related
how
differences
individual
Regulatorytamilies.
preschool
in
style approachesexamine
behaviour
in stressful situations.
children regulate affect and
(the
System
Coding
Narrative
MacArthur
The
makes use of a core set of stories
dealing
for
described
Battery
Stem
Story
with play
MacArthur
research groups
above)
is
between
the
facilitates
coding
system
and
groups
of
results
comparison
and
narratives
The
MNCS
themes,
three
parental
thesis.
content
areas:
assess
this
in
Lised
Content
themes.
themes
and
parental
representations
are
pertori-nance
and
i-eprescntations
different
coded
aspects
are
on
scales
of
performance
with
NvIlereas
categories
in
coded
57
defined scale points. The N1NCS may be sen as containing certain elements of both
In
representational and regulatory approaches. terms of a representational approach, the
for
the
to
thematic content and
story
stems
verbal
response
are
analysed
children's
is
Some
attention
also given to style of performance and type of
narrative coherency.
is
However,
differentiation
between
looking
there
no
expressed.
when
coding
affects
at
from
from
thematic
content
a
point of view or
a regulatory point of view.
affect
For
is
joy
laughs
the
coded
samewhen a subject child smiles or
or when the subject
example,
has the dolls laugh or say they are happy. Scales such as control and anxiety are more
behaviour.
to
to
the
subject's ability
regulate affect and
clearly related
2.5
THE LONDON PAR-ENT-CHILD
PROJECT
The empirical chapters of this thesis draw on data collected in the context of a
longitudinal study of attachment patterns being conducted by Drs. Howard and Miriam
College,
University
Centre
Freud
Anna
Fonagy
Peter
Professor
the
and
Steele and
of
Steele,
M..
Project
(Fonagy,
Parent-Child
London
known
The
London.
as the
studv is
Moran, Steele, H. & Higget, 1991; Steele, H, 1991; Steele, M, 1990).
launched
in
1987
Project
Parent-Child
London
first
The
when
was
phase of the
described
for
"a
the
hundred
studv
which
was
as
pregnant wornan were recruited
OIIC
how
bettcr
one's
own
experience
qf'childhood
influences
understanding
tit
allned
during
Recruitment
took
place
prenatal classes at
tile Inircnting qfthe next genet-ation".
58
the Obstetrics and Gynaecology Department of University College Hospital durim-, the
Selection
1987.
of
criteria included primiparous status. current
autumn and winter
father
fluency
in
English
language
the
the
the
of
with
child,
cohabitation
and ageabove
20 years. About 50% of those who were told about the study agreedto participate. While
last
information.
their
trimester
to
of
pregnancy
various
measures
in
were used collect
Demographic, personality, social context and attachment variables were assessed.
At approximately one year post-parturn, the second phase of the project began.
This included assessingthe twelve month old infant's quality of attachment with mother
Situation
The
Strange
Situation
father
Strange
Test.
the
was conducted again with
using
when the infants were approximately eighteen months old.
The third phase of the Project has now been completed. Table 2.1 (at the end of
instruments
including
Project
the
the
the
three
main
the chapter) outlines
of
phases of
Project
Parent-Child
London
The
two
the
third
consisted
of
visits
of
phase
assessment.
by the families. The first visit took place as close to the child's fifth birthday as possible
first
later.
For
their
the
the
and
parents were
children
visit
the
second six months
and
located
Centre.
The
Freud
The
Anna
based
laboratory
was
playroom
at
invited to a new
back
hut
to
built
the
the
one-way
mirror
enable
a
with
equipped
garden
in
in a purpose
first
in
brightly
Each
this
be
seen
mother and child pair was
tasks to
video-taped.
Chapter
Strange
Situation
6)
(described
the
modified
in
was
coloured playroom where
building
Following
taken
to
this,
the
the
parent was
main
yards a,,\-aN,in
administered.
Interview
E.
Attachment
Child
I)
Adult
(Appendix
for
lier
the
to
an
and
complete
ordcr
Beliaviour Checklist (Appendix Q. The child remained in the playroom and was
59
female
to
examiner who administered the MacArthur Story Stem Batter,
introduced
a
Story
The
Bus
tasks.
the
second visit took place in a new 'strange' place at
and
University College, London. In the second visit a modified Strange Situation with father
was conducted after which the child participated in other tasks with a female examiner.
2.6
CONCLUSIONS
The aim of this chapter was to review studies that have made attempts to assess
the internal world of the young child and to introduce the design and main instruments
This
this
study will explore the results of the application of the narrative
study.
used in
task to a large group of non-clinical five-year-old children. There also remains the need
for the MacArthur Story Stem Battery to have its psychometric properties investigated
through systernatised coding and inter-rater reliability analysis and this is addressedin
however,
has
begun,
Woolgar
(1996)
Work
following
this
the
was only able
on
chapter.
MacArthur
Narrative
the
the
to utillse
content and parental representation portions of
Coding System. This study is distinct in that it makes use of the entire MacArthur
Narrative Coding System, including performance codes.
Chapter Five presentsthe results of attempts to repli cate studies referred to above
between
the
tbund
children's responsesto the story stem task and the
associations
which
60
Checklist.
Subsequent
Behaviour
Child
chapters present associations
well established
to demographic variables collected before the birth and during the infancy of the target
children.
Later chapters report on the longitudinal and concurrent associationsbetween the

infant-parent,
child-parent and parental representations of
narratives with categories of
is
There
a relative absenceof work comparIng the MacArthur Stor-Y
attachment security.
Stem Battery responses to children's attachment history and this is the central aim
is
In
the
current
project.
addition,
an
attempt
made to construct theoretical
informing
insecure
battery
to
the
story stem
and to apply
profiles of secureand
these profiles to the four groups of attachment patterns in the sample.
'C
7,3
V
m
r-
a)
L-
L
-
-;r-
"17) cz
00 "
,
::
C: .M
V >,- C/)
-11:
E
0
m
to
cz
-=
m
.-
(D
oo
"
-1
0
.-
7z
cz
-=
m --d
--
=
Lr,
cz
ca
7D
Cc
L,
C)
=;
r,
j
.
17-
ct
"
Q)
t4
--
'a
12
4-
1*
4j
-I-- e:,
>
=
E --a
Q) - 0
m
=
"
2 C/
v
.
i2 c 2
"I
E
-c;
0
=
- 0 .--7 +
s.- "
;,,.2 V C/) C)
>
0
m
ct
-
v := m,
S
-- m .
S-
v 0
cz
;:z
5:
cz
"
cz
(1)
77-
&-
on
- CZI
-m
t' E v
0
z
"
(1)
V)
>
V)
E
M -a
-- .. CZ-p3 " -0
V) acz
(1)
) Z"Z:
. -
I-
J --
-0
cr
C'A
4-
(U
E a
4C3
>
t: CAU
cl,
M
V) a
Ln 0
La)
a
Ln
,)
31-
4(5
.<
4-
&(1)
M
2
E
U (D
=3
.t
(., V)
(A
-1 cz
CD-
CZ
U)
Ln
00
ct
V
V)
V
I=
C,
f4j 4t4
:z
z
-C
4--
ct
"
:
:
tz
-C)
U
Lf)
U
LV)
(: )
-10
a)
E
4-
r-
0,0
:
t.
E
_S5
C) 3-)
C ,
(1)
:)
.
>
CZ
L-
>, -- L 0
2
b
C.) -0
!t m
V)
Ln
f)
- ;::-
!.)
li
-=
L-
C:
n *= V 0
L.
L-
a)
;:-
C7
cz: t:
CL.
.=
0.
,: -C
! )
(.j
.)
4j V)
t
I--
>
CU .
5-
u U ICo
(1) cz
0
U
a)
V)
M.
> C) L
cz
>
-Z
=
.-
C)
'u "10 '

n :- )
I-)
CD
.
t:cz
cn
CL
4-
vV)
(A V) :
CA
M C) I-) (I)
.
VD
"10
2) E
ca.0
(1)
$a.
crj
4-
cL.
CA
7D
a.)
ci
4(A
V)
lll
C)
00
V)
bo
a)
5
lz
kn
l CIO
.
(I.)
(1)
&a)
C"j
.1
U"
62
CHAPTER THREE
ANALYSIS
OF THE MACARTHUR
STORY STEM BATTERY NARRATIVES
63
3.1 INTRODUCTION
This chapter is concerned with the ratings derived from the application of the
MacArthur Narrative Coding Manual (Robinson, Mantz-Simmons, Macfie, & The
MacArthur Narrative Working Group, 1995) to the story completion task obtained from
89 five year old children described in the previous chapter. The focus of this chapter is
derived
from
the
to
the children's story completions.
extent
which
reliable codes were
upon
The current study makes use of both audio and visual material from video
Thus,
transcripts
the
and
verbatim
of
children's stories.
it was
recordings, audio recordings
but
the way in which the
the
the
to
content of
children's stories
code not only
possible
both
themselves and their stories.
children presented
3.2 METHOD
3.2.1 Design
The subjects took part in a story stem
completion task
doll
play.
using
Battery
Stem
Story
MacArthur
known
the
These story stems,
as
hige Bretherton,
Helen Buchsbaum
Group are designed to elicit young

.,
developed by
Robert Emde and the MacArthur Narrative

children's
mental
Oppenheim.
(Bretherton,
technique
representations utilizing a narrative or storytelling
Eleven
Group,
1990).
Narrative
MacArthur
The
&
Enide
of these storv stems
Buchsbaum.
life
day
deals
to
typical
the
of
clay
a
in
relevant
an
ISSLIe
with
story
,,verc chosen and each
64
five year old child. For example, loyalty conflicts between a friend and a sibling.
brief
from
injuries
discipline,
familY
separations
parents, minor
and
punishment and
by
All
the
to
the
the experimenter who.
story
stems
were presented
children
relationships.
figures,
introduced
dilemma.
doll
The
the
the
the
narrative
containing
specific stor
using
in
dilemmas
Section
3.2.2.
When
their
the subject was male, the
are
outlined
and
stems
in
dolls
the
the
stories
children
were presented as male and male
were used,
charactersof
female,
female
dolls
The
the
stories were
and characters were named.
subject was
where
female
in
The
the
to
the
experimenter then asked the
same order.
children
presented all
her
happened
If
telling
the
to
what
next.
showing
as
well
as
narrative,
complete
children
the subject did not address the dilemma, the experimenter intervened with non-directive
Hence,
interviews
in
each
with children.
prompts similar to the techniques used clinical
form
to
of
complete, containing some
child was presented with eleven story stems
both
their
dilemma,
narrative.
actions
in
and
words
using
relationship
3.2.2 The story stem battery
full
for
I
A.
Appendix
below
(
the
protocol)
The eleven stories are surnmarised
see
issues
the
dilemmas
are
expected
children
the
which
about
explicit
and
the
embedded
vith
to elaborate.
I.
dinner
family
While
table. the protagonist
the
(SJ):
juice
the
ispilt
is seated at
discipline.
between
floor
juice
his/her
(issue:
the
and
nurturance
on
spills
accidentally
?
?
).
Does
Is
the
anyone clean up
mess'? anyone punished
What happens about
65
2.
Mother's headache (MH): The protagonist and mother doll are sittln,(,:
) together
watching television. The mother announcesthat she has a headache.tums off the television
find
to
the
and asks
child
something quiet to do on his/her own. A same sex friend of the
implores
the protagonist to let them watch television together (issue:
arrives
and
child
between
the mother's request and the friend's desire to watch television).
conflict
3.
Three's a crowd (3C): The protagonist and the same sex friend are playing ball
together in the garden, while the parents are talking to the neighbours at the other end of
the garden. When the younger sibling asks to join the game, the protagonist agreesbut the
friend refuses saying 'If you let your little brother/sister play, I won't be your friend
interests/loyalty
(issue:
between
the friend and the sibling).
anymore'
conflict of
4.
Burnt hand (13H): The mother is cooking dinner while the protagonist watches.
Father and sibling are sitting at the table. Mother warns the protagonist to stay away from
the stove but the child character cannot wait and knocks the pan off the stove burning
his/her hand (issue: direct disobedience has resulted in injury which requires attention).
Lost keys (LK): The parental charactersare involved in a heated argument in front
how
keys
lost
(issue:
to cope with parental conflict).
the
a set of
of
protagonist about who
6.
S-*Neetshop (SS): \N"hilein a sweet shop the protagonist asks the mother for some
he/she
has
had
day.
When
but
that
saying
already
refuses
some
she
mother is not
swects
looking the child character takes some off the shelf but
by
the shopkeeper
observed
fear/shame
disobedience
being
direct
by
ofinother
and
of
(issue:
caught
shopkeeper).
66
7.
Separation (SN): The parents leave on an overnight trip leaving the protagonist
(issue:
the
grandmother
separation anxiety).
and sibling with
8.
Reunion (RN): The grandmother announcesto the children that their parents have
following
home
(issue:
behaviour).
the
morning
arrived
reunion
9.
Bathroom shelf (BS): The mother leaves the house briefly, having prohibited the
from
touching anything on the bathroom shelf. During play the younger sibling
children
finger
their
and asks the protagonist to get a plaster (off the bathroom
accidentally cuts
The
but
the
the
the younger
sibling
of
mother's
protagonist reminds
prohibition
shelf).
between
his
finger
bleeding
indicates
(dilemma:
the
conflict
aiding the sibling
is
sibling
and obeying the mother's prohibition).
10.
Exclusion (EX): The protagonist's parentsare sitting on the couch and the samesex
have
time
together.
the
to
their
the
to
some
alone
can
so
parents
room
go
child
parent asks
The child goes to their room and the experimenter shows that the same sex parent then
leans over and gives the other parent a hug (issue: oedipal conflict).
11.
Biscuit tin (13T): The younger sibling takes a biscuit from the tin and the
biscuits'.
The
'No
him/her
then
the
that
sibling
said
more
mother
reminds
prota(.,onist
does
(issue:
then
tell
their
the
the
who
enter
that
not
parents
will
protagonist
pleads
t
Loyalty
'?
the
to
tattic
siblino
sibling or parental rules).
on
protagonist
67
3.2.3 Subjects
The 100 families who participated in PhasesI and 11of the London Parent-Child
Project were contacted shortly before the children's fifth birthday and asked to participate
Attrition
12
isit
follow-up
6
%)
7
the
(or
the
since
research.
month
vi
I
was
children
and
in
(or
10
10
%).
One
18
the
children
was
child, whose parents participated
month visit
since
joined
for
but
for
12
18
the
then
the
and
month visits
cohort
pre-natally, was unavailable
the 5 year assessment.The single greatestidentifiable cause of attrition being the fact that
the families had moved outside of the UK.
Ninety-two of the original families agreed to participate. The mean age of the
Thus,
59-65
(range
61
boys)
49
(43
ninety-two
months).
months
was
girls and
children
in
took
part the study, with eighty-nine actually completing all eleven story stems.
children
difficulties
due
to
technical
with
It was not possible to code two of the children's stories
battery
but
began
One
to
the
the
tape.
refused
child
either the video or audio portions of
between
took
that
testing
The
procedure
a
of
as
part
administered
were
stems
complete it.
in
hours
the
half
the
parents.
two
of
absence
and
and
one
3.2.4 Procedure
lasting
latter
tasks
the
The story stems were presented in
part of a series of
following
half
hour
the
the
strange
modified
of
administration
and
a
an
approximately
4-female
A
five
the
to
the
t'()r
mother.
introduced
with
experimenter
vas
olds
\-cars
,
situation
68
left
them in the testing area so she could also 'tell stories' (the
then
the
subject and
mother
AAI) to another experimenter. All the children's stories were presentedin the playroom of
the Anna Freud Centre which is equipped with a one-way mirror to enable the procedure
to be unobtrusively video taped. The subjects were also audio taped using a Sony
Professional Walkman.
The subject was introduced to the Duplo doll characters; the protagonist, a same
doll
(called
George
for
boys
Susan
for
age
child
same
sex,
and
girls), a same sex younger
for
boys
for
(Bob
Jane
and
sibling
girls), a same sex, same age friend (Dave for boys and
Laura for girls), a mother, a father, a grandmother and a shopkeeper.In addition there were
Duplo
of
props such as a table, chairs, sofa, a television, a cooker, and a
a number
bathroom set to facilitate the story telling (seeAppendix A. I for the list of props by story).
The experimenter began the standard stem for each story specifying the use of both doll
invited
by
'You
the
the
to
then
the
child
complete
story
saying,
props and
characters and
behavIOUral
The
happens
tell
subject'sspontaneousverbal and
next'.
me what
show me and
if
did
followed
by
the
particularly
child
standard, non-directive prompts,
responseswere
not addressthe conflict in the stems.
For example, the spilled juice story contains a dilemma in which the focus child,
in reaching for some juice, spills the juice all over the floor. This stem was used to
Duplo
dolls
The
to
the
task.
the
to
and
child was encouraged place
introduce the subject
before
table
the
the
a
presentation
of
around
stem
seated
their
positions,
initial
props in
foster
dolls
In
familiarize
the
the
them
and
participation.
spilled juice
to
with
was started,
dilemma
bY
the
the
moving
characters
and
props
emphasised
the
experimenter
stem,
69
leading
including
to
the
the act of spilling,
through the sequence of events
accident
up
before
In
itself
to
the
the
this instance the
request
complete
stem.
emphasised,
which was
juice
floor;
the
the
spilled on
standard prompt concerned
'What happens to the juice
George/Susanspilled on the floor T and was asked either when the subject did not respond
dilemma.
did
The
the
their
not address
experimenter articulated the subject's
response
or
for
behavioural
the
the purpose
responsesand repeated
subject's vocalizations
character's
further
by
At
the
the end of each stem,
to
encourage
elaboration
child.
and
of clarification
the experimenter immediately set up the props for the next story stern involving the suject
both
The
the
the
moving
experimenter
stems involved
presentation of
as much as possible.
dilemma
leading
to
the
she was
up
characters and props through the sequenceof events
introducing.
forty-five
battery
took
the
Typically the administration of
approximately
entire
from
tapes
the
and
video
transcribed
The
audio
verbatim
were
responses
subjects
minutes.
tapes where necessary.
Scheme
Coding
Narrative
MacArthur
later
the
The story stems were
coded using
Group,
1992).
Working
Narrative
MacArthur
&
The
Macfie
Mantz-Simmons,
(Robinson,
for
System
Coding
this
Narrative
MacArthur
chosen
Two,
was
Chapter
the
As discussedin
for
basis
a
common
represent
which
stories
because
of
set
core
a
of
it makes use
study
between
facilitates
groups.
of
results
comparisons
dealino
and
narratives
with play
groups
2.
A.
Appendix
Fhe full rating protocol is provided in
by
developed
JoAnn
Manual
Coding
Scheme
Narrative
was
The MacArthur
70
Robinson, Linda Mantz-Simmons, Jenny Mac Fie and the MacArthur Narrative Working
designed
for
This
manual was
use with the MacArthur Story Stem Battery and
group.
attempts to quantify the children's responses to those story stems. The system xvas
designed to be coded from video tape only, but the availability of the transcripts from the
facilitated
the coding, and so both media were used.
tapes
audio
During the first part of 1995, the MacArthur group produced the coding manual and
for
first
training
the
time in Europe at University College, London.
course
presented a
Three of the four raters, including the author, attendeda week long workshop led by Linda
Mantz-Simmons
Psychoanalysis
Unit at University College,
took
the
which
place in
London. The purpose of the workshop was to train a group of students, mental health
in
Narrative
Coding
MacArthur
Scheme.
the
professionals and psychoanalytic researchers
The week of training was organised in terms of watching videos of children filmed during
the story stem administration and discussing, as a group, the specific content and
determining
Important
to
this
themes
process was
what particular
observed.
performance
behaviouralresponsesof the children would be coded. For example,it was agreedthat a

be
the
rated as concern
examiner would
child's raised eyebrows and visual referencing of
distress.
be
Minor
by
fear
the
coded
as
child would
exhibited or expressed
and instancesof
MacArthur
this
time
the
to
at
and are
were
made
manual
coding
changes and additions
fourth
The
2.
the
A.
the
trained
Appendix
three
workshop
raters who attended
in
noted
hours
forty
four
together
to
These
then
as
a
sixty
group,
spent
an
additional
raters
rater.
tapes
their
skills.
and
coding
III
transcripts.
video
refining
watching
readino
Z..
The
The Nlac.Ailhur Narrative CodingL-Manual is divided into th
main
sections.
-ree
71
first deals with content themes, the second with parental representationsand the third Nvith
include
Content
themes
themes such as aggression, affection injury
performance codes.
,
Parental
describe
representations
and punishment.
references made bY the children to the
parental characters and are specified as positive, negative and disciplining. The
performance of the narrative includes such scales as affects expressed, style of'
be
performance, relationship to the examiner and narrative coherence.This studN.,,
will
one
first
developed
the
to
the
of
apply
recently
coding scheme to a group of non-clinical five
year old children whose attachment status at infancy is known.
3.3
RESULTS
3.3.1 Coding reliability
This section divides the analysis of the inter-rater reliability of the story stems
derived from the MacArthur Narrative Coding Manual into three sections as is the manual.
The first deals with content themes,,the secondwith parental representations and the third
include
Content
themes such as aggression, affection,
themes
with performance codes.
injury and punishment.
One of the performance variables, overall adaptiveness of response, is not part of

the original MacArthur Coding Manual and represents a theme added by Dr. Miriam
Steele, of the Anna Freud Centre, and Dr. Jill Hodges, of Great Ormond Street Hospital.
Fhis scale attempts to assessthe extent to which the children demonstrate, either through
they
the
that
the
understand
contlict
embedded
within
actions.
ston, stem.
or
words
72
Furthermore, it also attempts to assessthe quality of the resolution arrived at bv the
children to each particular conflict.
Content themes
Eight-six children gave responsesto the story stemsthat contained codeablecontent
themes.These stories were coded by the four raters who were divided into two independent
inter-rater
the
pairs and
reliability assessedwith Cohen's Kappa. As the range between the
two pairs of raters was so low, the median scores of the two pairs of raters are reported.
Six themes (see Table 3.1) could not have Kappas computed either becausethey
by
both
in
were not seen
raters each team (competition, sharing, and rivalry) or becausenot
levels
by
both
theme
the
all
of
were seen
raters in each team (shame, blame, and
repetition).
Table 3.1 : Median Cohen's Kappa scores for the inter-rater reliability of the 24 content
themes as presented in the manual.
Theme
Kappa
Theme
Competition
Shame
Sharing
Affiliation
Rivalry
Blame
Exclusion
Aggression
Empathy
Refused empathy
Interpersonal con.flict resolution
'Fattling
Compliance
Non-cOrnpliance
I-scalation of conflict
53
.
70
.
67
.
28
.
68
.
77
.
3)
.8
79
.
Affection
Teasing
Dishonesty
Punisliment
Reparation
Verbal conflict
Injury
Atypical response
Repetition
Kappa
35
.
82
.
27
.
72
.
67
.
71
.
28
.
58
.
79
.
/-i
The Kappas ranged between 21 - 83 Five themes presented in the manual as

.
.
(refused
empathy, escalation of conflict. affiliation. teasing and
codes
presence/absence
by
both
but
Exclusion
team
each
raters in
were not reliable.
was
verbal conflict) were seen
is
level
in
Tables
three
theme
as
when
rated
a
and
presented
only moderately reliable
-3.2
interpersonal
(aggression,
The
below.
33.33
themes
twelve
remaining
empath.y,
conflict
and
dishonesty,
tattling,
compliance, non-compliance, affection,
punishment,
resolution,,
Kappa
injury
733
of .
and atypical response) were reliably coded with a mean
reparation,
(range of. 58 to. 83). Repetition could not be reliably rated and was seen so few times bv
from
further
it
be
dropped
%
4.65
946
(44
time)
that
the
times in
will
the raters
of
stories;
analyses.
The complex themes (exclusion, aggression, empathy, interpersonal conflict

both
blame,
recoded
were
response)
atypical
and
reparation
punishment,
shame,
resolution,
derived
from
latter
The
(ie.,
binary
themes
present or absent) and as new variables.
as
into
for'both'
including
the
into
the
each
of
code
separatevariables,
splitting the categories
for
T
in
the
For
as
manual
operationalized
was
aggression
example,
new variables.
for
both.
Hence
the
for
'2'
of
new
variable
and3'
aggression
verbal
physical aggression.
'T
T
from
the
codes.
and
combining
t:
created
was
physical aggression
74
Table 3.2: Median Kappa scores for inter-rater reliability of the complex content themes
form
binarv
to
variables.
collansed
Binary Theme
Kappa
Binary Theme
Kappa
Exclusion
71
.
74
.
67
.
72
.
83
Shame
70
.
29
.
75
.
73
.
Aggression
Empathy
Interpersonal conflict resolution
Atypical response
Blame
Punishment
Reparation
Table 3.3 : Median Kappa scores for inter-rater reliability of the new variables derived
from the complex content themes.
Kappa
Theme
83
Self exclusion
Other exclusion
Physical aggression
Verbal aggression
Practical reparation
Guilt reparation
B ame other
Blame self
.
74
.
79
.
01
.
70
.
67
.
17
.
Theme
Shame other
Kappa
74
.
Shame self
Empathy
Helping
Physical punishment
Verbal punishment
Positive atypical response
Negative atypical response
6.33
.
00
.
84
.
62
.
67
.
89
.
The important result from Table 3.2 is that blame remains unreliable as a binary
divided
blame
Furthermore,
this
self
into
and other
code remains unreliable when
code.
blame (Table
from
further
dropped
Interestingly,
be
analyses.
shame
will
and
as
such
-3.3)
hence
by
(not
is
this
all
removing
unreliable component
raters),
seen
unreliable
also
self
fIrom the binary code increasesits reliability. Verbal aggressionis unreliable and so is the
definition
to
the
helpino variable which was added
of empathy/helping in the manual by
Table
33.3
Elevcn
(self
new
variables
in
sixteen
are
satisfactorily
ofthe
reliable
team.
the
acygression,
physical
practical
reparation,
I
exclusion.
guilt
reparation,
other
exclusion,
L-
75
shame other, empathy. physical punishment and verbal punishment. positive atypical
response and negative atypical response) with Kappa's ranging from 62 to 89.
.
.
Parental representations
The parental representations data was analysed for inter-rater reliability as above.
firstly as presented in the manual, then as binary codes and finally creating new variables
according to the representation of parental agency, either as mother or father. Table 3.4
for
first
the
the
two analysesand Table 3.5 the agencyreliability. Table 3.4
presents
results
disciplining
the
that
themes achieved satisfactory reliability as
shows
positive and
presented in the manual and also as binary codings, but that the negative representations
were not reliable.
Table 3.4 : Median inter-rater reliability for the parental representationsmanual categories
binary
the
codim,y and
codinv,.
Parental
Three by Three Kappa
Representations
Positive
Negative
Disciplining
65
.
25
.
72
.
Presence/Absence Kappa
67
.
29
.
74
.
Table 3.5 : Median inter-rater reliabilitv for mother and father representations.
Mother
POsltl%c
Negative
1
Disciplining
64
.
72
.
.72
Father
I1
61
.
76
.
68
.
Table 3.5 Licnimistratesthat by recoding the three-way codes. into those where the
is
directly
by
designation.
the
*both'
the
cithei
i
stated
or
IIimplied
parents
increases
"It-,clicy ot
76
the reliability of the negative representations substantially. The designation of agency in
the children's narratives can be problematic and given that the ratio of direct assignments
'both'
25
the
to
across
stories
was
of agency
negative/16 both for mothers and 1-1
for
fathers,
both
the negative parental representation code was considered
negative/16
decision
However
the
was made to retain the use of these codes for future
questionable.
analyses.
Performance themes
Eight-nine children were coded for performance themes.As for the content themes,
the performance section was coded by the two pairs of independentraters and the reliability
Spearman's
Cohen's
Kappa,
Pearson's
R
Rho
and
where appropriate.
was assessedwith
For easeof presentation the performance section is described in two groups, the
Performance
those
themes
themes which
themes.
the
themes
are
affect
and
perfon-nance
both
to
the story telling task
the
to
attempt assess child's overall perfon-nanceand attitude
(denial,
the
the
role of parents, narrative coherence,
examiner
and
relationship with
involvement
the
to
the
of
examiner, investment
examiner,
responsivity
style,
performance
in performance and the child's understanding of the conflict). These perforinance themes
The
by
task.
themes
the
the
three
the
affect
of
story
stem
over
all
phases
rater
arejudged
include scales that attempt to assessthe emotional openness of the child during the task
distress.
These
themes
anger.
concern,
sadness
and
anxiety).
(expi-cssionsof control, joy.
for
three
the
the
task.
phases
of
\\,ci-c coded separatcly
77
Cohen's Kappa and Pearson's R were calculated first for the performance themes
in
Table
3.6.
Two
themes, the role of parent (K = 82) and child's
and are presented
.
understanding of
the conflict (K =. 68), were reliably coded as complex themes as
The
the
manual.
remaining six performance codes (direct vs indirect
presented in
performance style, child responsivity to examiner, child involvement of examiner,
denial
and adaptivenessof response)yielded low to moderatc
investment in performance,
Pearson and Cohen's kappa scores.
Table 3.6 : Median inter-rater agreement for the performance codes as presented
in the manual.
Pearson's r
82
.
68
.
47
.
Kappa
% Agreement
82
.
68
.
42
.
88.21 %
87.92%
Child responsivity
to examiner
54
.
37
.
67.60%
Child involvement
of examiner
Investment in performance
55
.
47
.
67.50%
77
.
46
.
46
J .
52
.
43
.
32
..
57.52%
Rating Scale
Role of Parent
Child's understanding of conflict
Direct vs indirect performance
87.74%
style
Denial
_
jAdaptiveness
of response
76.05%
.
50.42%
Five of these themes (child's responsivity to the examiner, child's involvement of

denial
of
and
aclaptiveness
response) were recoded
performance,
in
investment
examiner,
Table
3.7.
in
as new variables and are presented
78
Table 3.7 : Median scores for inter-rater reliability of the complex performance
themes collar)sed to form new variables.
Rating Scale
Pearson's r
51
.
55
.
63
.
49
.
55
.
Child responsivity to examiner

Child involvement of examiner
Denial
lAdaptiveness
of response
,
Kappa
% Agre ment
77.25%
47
.
54
.
52
.
48
.
40
1.
76.45%
69.00%
80.01) %
62.3 3%
1
By collapsing the complex themes to form new variables, three of the five
scales'
(child responsivity to examiner, child involvement of examiner and denial) Kappa scores
improved yet remained moderate. Investment in performance remained the same(K =. 52).
Unlike the previous scales, narrative coherence is not a hierarchical scale but a
inter-rater
As
categorical one.
such
reliability was calculated using Spearinan's rho. It was
for
before
to
the
narrative coherence code
not possible calculate reliability
recoding this
levels
into
After
this
all
saw
all
of
scale.
raters
scale
a new variable as not
recoding into
in
Table
3.8.
to
the
reliability shown
a new variable it was possible calculate
Table 3.8 : Median inter-rater reliabilitv for narrative coherence.

Rating Scale
Narrative cohemice
Spearman's
.
50
rho
% Agreement
60.03 %
Although only moderate agreement was reached regarding some of the above
future
because
be
tor
the
the
anal.
ysis
of
of
retained
importance
thc'
nature of
will
scales,
be
discussed
Difficultics
stions
concerning
and SLI,
reliability will
(',,
(,Ie,,
at the end
the scales.
of this section.
79
Affect themes
Table 3.9 presentsthe Pearson's r and Cohen's kappa for the affect themes in
each
of the three phases of the story stem completion task. Kappas could not be calculated for
six of the codes (anger in the presentation and transition phase, distress in the transitioii
in
phase,concern the narrative phaseand sadnessin the presentation and transition phases)
in
they
as
are presented the manual either becausethey were not seenby both raters in each
team or because not all levels of the theme were seen by both raters. The percentageof
between
the raters for the majority of the scales is quite high.
agreement
Table
Median
for
the affect themes as mesented in the manual.
:
inter-rater
reliabilitv
_3.9
Rating Scale
Pearson's r
Kappa
% Agreement
Control (p)
58
.
52
.
63
.
58
.
62
.
56
.
55
.
47
.
59
.
46
.
45
.
45
.
89.06%
Control (n)
Control (t)
Joy (P)
Joy (n)
Joy (t)
66
.
55
.
Distress (n)
61
.
58
.
58
.
61
.
52
.
48
.
85
83'
.
36
38
.
Anxiet\ (11)
Note:
99.38%
95.27 1
88.07%
99.79%
96.86%
99.88%
Sadness (t)
Anxict (p)
87.04%
99.71 %
Sadness (p)
Sadness (n)
72.62%
97.93%
Concerti (n)
Concern (0
63.03 %
99.56%
Distress (t)
Concern (p)
64.33 %
99.68%
Anger (t)
Distress (p)
91.58%
97.57%
Anger (p)
Ans4er(n)
89.47%
49
.
49
.
p= presentation phase
46
.
48
.
n= narrative phase
74.52%
77.46 qO
t= transition phase
80
As can be seen in the table above, for some of the scales, concern in the transitioti
phase for example, the Kappa is high and the percentage of agreement is correspondi ngly
high. For others where the Kappa is moderately high or not able to be
calculated because
the frequencies of occurance are too low, the percentage of agreement still remains high.
This is because,although the frequencies were low, the raters did
large
agree a
percentage
few
the
time
the
times they did see the theme. For example, distress in the transition
of
phase appeared twice and the raters agreed 99.56 % of the time. These scales were
fonn
binary
to
collapsed
variables and the results are presentedin Table 3.10. Control and
anxiety are already binary codes as Presentedin the manual.
Table 3.10 : Median inter-rater reliability of the complex affect themes collapsed to form
binarv variables.
Rating SCale
Joy (P)
Joy (n)
Joy (t)
Anger (p)
Anger (n)
Anger (t)
Distress (p)
Distress (n)
Distress (t)
Concern (p)
Concern (n)
Concern (t)
Pearson's r
49
.
55
.
52
.
23
.
64
.
81
.
61
.
67
.
50
.
50
.
13
.
85
.
Sad (p)
Sad (n)
Sad (t)
Note:
Kappa
% Agreement
48
.
55
.
50
.
22
.
64
.
79
.
58
.
64
.
50
.
49
.
15
.
3)
.8
74.23
p= presentation phase
77.88
98.82
90.48
99.90
99.38
95.75
99.67
88.42
98.05
99.89
99.71
41
i.
*I
84.89
338
.
*
n= narrative phase
96.87
99.88
t= transition phase
frequencies
low
BCCLILI,
the
of the affect variables within the separatephases
SCof
81
it was decided to combine the three phases of the narrative. In combininu, the scales to
form binary variables the nature of the data changed and it was no longer possible to use
Cohen's Kappa and so Pearson correlations were used.
Table 3.11: Median inter-rater reliability of affect themes when three phases of the
narraLlVe LaSKare comoineu.
Rating Scale
Control
Joy
Anger
Distress
Concern
Sadness
Unxiety
Pearson's r
69
.
67
.
65
.
58
.
52
.
30
.
57
.
As can be seen,collapsing the three phasesgenerally improved the reliability and

further analyses will utilise the affect codes combined across the three phases of the
be
dropped
However,
theme,
sadness,was still not reliably rated and so will
one
narrative.
from further analysis.
3.4
DISCUSSION OF THE RELIABILITY
RESULTS
failure
for
of obtaining
successand
This section will explore the possible reasons
the
themes
and
themes.
perfon-nance
and
representations
parental
the
content
reliability on
for
moderate
and
good.
proved
poor.
reliability
which
divided
those
areas
into
is
82
Unreliable themes
First considered are those themes that were not reliably rated by the coders. Eight
in
binary
be
the
themes
to
themes
manual
as
presented
proved
unreliably coded.
content
The first three content themes (competition, sharing and rivalry), could not have kappa
binary
by
both
themes,
this
that
they
teams of
means
were not seen
calculated and, as
for
low,
is
hardly
frequency
The
these
each of
was
which
surprising given that the
raters.
battery
designed
in
MacArthur
to
the
the
code
entire
range of stems
wliicli
manual was
included
in
battery.
issues
directly
but
this
these
were not
includes stories which address
Nonetheless, spontaneousintroductions of thesethemeswould be coded but the occurances
Similarly,
low
these
and unreliable.
were
of
the other five themes (refused empatliy,
frequencies.
low
had
very
escalation of conflict, affiliation, teasing and verbal conflict)
This is a problem with the kappa statistic which is very sensitive to differences when the
(15
times out of
low.
Escalation
base rates are extremely
of conflict was also rarely seen
946 stories) and was also difficult to code because the operational isation required a
by
into
the
the
level
initial
subject.
stem
introduced
the
of
conflict
judgement about
left
i
but
the
the
interesting
sation
operational
again
code
Affiliation is a potentially
positive
had
That
to
the
too
make
a
coder
is,
open-ended.
the
of
inclusion
sense
judgement about
do
that
that
those
must
give
or
an
activity
in
that
participates
everyone
either
judgement
Similarly,
belonging".
the
code
conflict
verbal
-a
or
of
inclusion
sense
the impression of
beyond
inflected
"highly
extended
remarks"
angry
verbal
of
the
identification
required
Part
dialogue.
the
this
the
some
was
pleasure
with
instance
of
problem
a
into
and
a sim3le
'udgements
in
of
with
reliable
interfering
ar(gurnents,
take
representing
children \VOLIld
Ii
. continued angry inflection'.
83
The themes of shame and blame were not binary themes and as such the inability
to calculate the kappa was due to at least one of the levels not being rated by one of the
teams. As Table 3.3 above indicates, blame proved to be unreliabl.y rated both as a binary
broken
down into the two categoriesof blame other and blame self Shame
code and when
proved somewhat more complicated in that it was reliably rated as a binary variable,
however, when broken into the two categories of shame other and shame
self only shame
self was not reliably rateable.
Only one of the performance themes, an affect code, proved to be unreliable

(having Kappa scores of below 40) when recoded as a new variable. The affect thernes
.
difficult
to rate consistently well when divided into the three phasesof the
proved more
be
due
fact
This
to
the
that decisions about when the three phases
task.
story stem
may
began and ended were often difficult to judge, particularly judging when the narrative
began.
frequencies
The
instances
the
transition
phase
of
of anger,
phase ended and
distress, concern and sadnessin the three separatephases was also quite low. This could
be due to the non-clinical nature of the sample children. Anger in the presentation and
transition phase,distress in the transition phase,concern in the narrative phaseand sadness
have
Kappas
Yet
transition
the
the
could
not
calculated.
phases
and
presentation
in
being
99.05
%
for
themes
these
present was a mean of
when the theme was seen
agreement
by the raters. Only sadnessproved to be unreliably rated both when considered in each of
the three phases and when the three phases of the task %verecombined. As such sadness
further
from
dropped
be
analyses.
will
84
Moderately reliable
There are many scalesthat can be considered to be moderatelY reliably rated, that
is, having obtained Kappa scoresof between 40 and 60. One performance scale presented
.
.
(performance
binary
the
manual
code
in
style) was rated moderately reliable (K = .4-1).
as a
Five of the performance scales proved to be moderately reliable when recoded into new
involvement
(child
to
responsivity
examiner, child
of examiner, investment in
variables
denial
Kappa
48
(range
and
adaptiveness
of
response)
with
a
mean
of .
performance,
of .40
For
54).
to
the
the
child's
example,
responsivity
examiner
is
presented
manual
as
in
a
from
being
the
child
unresponsive to the examiner to the child
seven point scale ranging
When
the
to
rated as a seven point scale the
examiner.
responding enthusiastically
into
distinguishino
divided
When
three
this
a
point scale
code is
reliability is poor.
r-between no response,reluctant responseand ready or enthusiastic responsethe reliability
i
level.
The
thi
the
coders
s
scale
encouraged
to
sation
of
I
operationall
is increases a moderate
distinguish
did
this
not clearly
category which
to adopt a 'holistic approach' when coding
between the different levels of response. Difficulties with the coding of this scale might
differences
between
to
this
due
be
assess
to
attempting
is
to
code
whether
as
confusion
also
to
the
the
task
examiner.
responding
the
subject
to
or
story stem
the subject responding
Three
be
injury,
to
of the
One content theme, that of
moderately reliable.
proved
the
three
be
of
phases
across
collapsed
I
to
when
fect
reliable
moderately
themes
proved
at
(distress,
task
concern and anxiety)
the narrative
be
This
to
code is quite an
Narrati%c coherence also proved
moderatek, reliable.
tI
the
perspective
communication
on
attachment
of
a
N,
iex%rom
particularly
one.
important
85
theory and cognitive theory on coherence,and problems with the operationali sation of this
be
because
Difficulties
to
the scale attempts to addresstwo
addressed.
arise
scale needs
issues
how
that
the
the
the conflict in the stor
at
once,
of
coherence
of
narati-,.,
main
eand
is addressedand the type of resolution that has or has not been reachedby the child. Von
Klitzing, Kelsay and Emde (1998) have noted similar difficulties and in responseto this
by
the
two
aspectsapart transforming the narative coherencecode into nex\
problem pulled
This
Conflict'
'Coherence'.
'Embellishment/Addressing
the
solution seems
and
scales of
highly reasonable and the two new scales are recommended for future use of the
MacArthur Narrative Coding System.
Adaptiveness of response, which attempts to make subjective assessments

It
much
the
is
reliable.
moderately
also
was
the
of
response,
adaptiveness
overall
regarding
of
difficult
assessments
with
impressionistic
agreement
to
inter-rater
good
obtain
more
bad
or
average
response
because
a
good,
constitutes
of
what
judgements
value
responses
is required.
Reliable themes
in
(exclusion,
the
manual
Twelve of the content themes were reliable as presented
conflict
interpersonal
empathy,
aggression,
resolution,, tattling,
compliance, non-
dishonesty,
atypical
and
response).
reparation
punishment,
affection.
compliance.
broken
into simpler yet more informative scales a total of seventeen
Furthermore, when
(self
physical
a(ygiression.
other
exclusion,
exclusion,
reliable
proved
content %ariables
physical
punishment,
other,
empathy,
verbal
shame
reparation,
guilt
practical reparation.
86
punishment, interpersonal conflict resolution, tattling. compliance, non-compliance.
dishonesty,
affection,
positive and negative atypical response).
When looking at the performance themes two (role of parent and child's
understanding of conflict) were highly reliable as presentedin the manual. When the affect
themes were collapsed across all three phasesof the narrative task three (control, joy and
anger) proved highly reliable having kappas that ranged from 65 to 69.
.
.
Parental representationsproved reliable when broken down into which parent was
disciplining.
decided
for
It
as
represented positive, negative and
was
easeof analysis and
interpretation to use the parental representationsvariables that proved reliable when coded
for
for
father
in
each of the three categories,that is positive, negative
separately mother and
disciplining
categories.
and
In attempting to develop a systematic and empirically useful coding scheme for

findings
(1991)
interviews
Marans
that
to
the
pertain
et al
raised several points
clinical
described above. On the positive side, even with the complexity and variability of a play
interview, agreement as to what general themes the child is conveying can be reached.
However, they emphasisethat those raters who shared a common theoretical perspective
On
have
higher
the
they
tended
to
task
their
to
more
negative
side,
agreement.
also
relevant
found that the nature of the operational definitions of the thematic categories sometimes
judgements
inferential
less
for
and
agreement among raters.
more
allow
has
been
by the rating team and that is that the
also
problem
identified
-\nother
87
item
(Woolgar,
1996).
For
to
the
meaning of an
may change according
story
example, a
in
'Exclusion'
the
and 'Headache' stories may represent a pro-social
code of compliance
indicating
empathy and respect toward the parents. In contrast, compliance in
action
'Bathroom Shelf represents an obedient but not necessarily an empathic way of solving
for
dilemma.
has
This
the making of scalesacrossthe stories
problem
a moral
implications
is
and addressedin the next section.
3.5
REDUCTION
OF DATA
The need to create meaningful aggregatesout of the large number of single coding
Warren
(1996)
For
been
has
et
al
example,
taken
studies.
some
in
account
into
categories
developed a coding scheme which focused on children's emotionality and selfdata
the
(1996)
Woolgar's
and
parental
content
only
used
of
reduction
representations.
important
include
did
the
and
performance
potentially
not
representation variables and
affect variables.
3.5.1
Scaling
data
to
the
To address the need reduce
and create reliable aggregates out of the
from
be
Ie
themes
the
section
categories,
previous
wlll
coding
siii-(-,
of
number
basis
high
levels
investigation
further
the
on
t'or
of
acceptablY
of
inter-rater
considered
88
having
been
justif,
In
to
achieved.
order
reliability
collapsing the themes acrossthe elevei-i
1
3.1
in
below
Tables
the
and
alpha scores were calculated and are presented
stories
).
-1.3 -3
3.14.
Table 3.12 : Reliable content codes and their aggregatealphas across the II stories
Theme
Theme
Interpersonal conflict resolution
Tattling
Compliance
Injury
Self exclusion
Other exclusion
Physical aggression
19
.
09
.
19
.
13
.
30
.
62
.
62
.
Empathy
Practical reparation
1.21
Non-compliance
Affection
Dishonesty
Shame other
Guilt reparation
Physical punishment
Verbal punishment
1
Positive atypic
09
.
44
.
52
.
-55
56
.
66
.
54
.
61
.
1.67
have.
differenct
theme
the
may
meaning a
As mentioned, problems concerning
Compliance
least
themes
internal
two
and
above.
have an effect on the
consistency of at
being
in
depending
they
different
have
are
story
which
meanings
very
non-compiance can
used.
from
derived
the content codes
In total, eight of the aggregate variables
injury,
tattling,
selfcompliance,
(interpersonal conflict resolution, non-compliance,
higher
40
did
than
alpha
scores
achieve
not
reparation)
practical
and
exclusion, empathy
.
factor
in
included
be
the
analysis.
not
will
and so
Table 3.13 presents the parental representaions and their aggregate alpha scores.
89
Table 3.13 : Parental rer)resentationsand their avgregate alphas across stories
Theme
Theme
Mother
Father
Positive
Negative
.
Disciplining
Positive
48
.
25
.
49
.
Negative
Disciplining
29
.
15
.
41
..
Of the parental representation codes half failed to score an alpha higher than 40.
.
Only positive and disciplining
disciplining
paternal
maternal representations and
One
factor
included
be
be
the
that
the
therefore
of
reasons
analysis.
in
will
representations
fact
due
be
failed
father
the
to
that
to
two of the three
achieve consistency may
codes
Father
is
the
a
story
appears
in
never
the
stems.
story
character
in
parental
main
mother
in
figure
is
the
and
is
not
stories
of
seven
the
a
peripheral
character,
mother
without
in
mentioned three of the stems at all.
Table 3.14 presents the performance and affect codes and their aggregate alpha
scores across stories.
90
I-able 3.14 : Reliable performance and affect codes and their aggyregatealphas across
stories
Theme
Theme
60
.
58
.
_
Indirect vs direct performance style 96
.
_
96
Child's responsivity to examiner
.
79
Child's involvement of examiner
.
_
91
.
43
Denial
.
Narrative coherence
Control
Role of parent
Joy
Anger
Distress
Concern
Anxiety
67
.
94
.
92
.
61
.
55
.
77
.
9-1)
.
Of the performance and affect codes all their aggregate scores across the eleven
fact
by
is
96.
This
from
43
40
that
the
explained
ranging
above
stories scored
- .
.
.
dependent
the
type
themes
of story stem preented
upon
as
are
not
performance and affect
as the content and parental representaion codes are.
3.5.2
Factor analysis
in
be
to
related one another and
Many of these single scales created above may
factor
data
further
strategy
task
analytic
the
a
tackle
reduction,
of
this
and
order to address
be
both
to
The
and
that
rated
reliably
twenty-seven
proved
be
scales
adopted.
,,vill
factors
The
factor
together.
sorted
analysed
were
stories
eleven
all
consistent across
in
below
Tables
loadings for both the unrotated and varimax rotation analyses are shown
0.50
loaded
Variables
and
abo-ve
are
of
absolute
an
value
3.16.
with
which
and
reported.
91
Table 3.15 : Sorted factor loadings for the unrotated solution
Factor I
Factor 2
Factor 3
Factor 4
61
.
72
.
74
.
70
.
Role of parent
Responsivity to examiner
coherence
_Narrative
I
style
understanding of conflict
_Child's
Physical aggression
67
.
61
.
56
-.
Anxiety
Disciplining father
54
.
54
-.
Positive maternal representation

Verbal punishment
Disciplini
50
.
g mother
60
.
56
-.
65
.
Affection
1
67
.
Table 3.16 : Sorted factor loadings for the varimax rotation

Factor I
Role of parent
ResponsiVity to examiner
Narrative coherence
Joy
style
_
Excludes other
Physical punishment
Shame other
Verbal punishment
Disciplining father
Disciplining mother
Physical aggression
Control
Anxietv
_
Denial
_
Affection
Factor 2
Factor 3
Factor 4
54
.
78
.
72
.
62
.
74
.
73
.
57
.
53
.
56
.
74
.
68
.
78
.
50
.
66
.
68
.
64
.
533
-.
54
-.
73
.
82
.
92
The varimax rotation is preferred becauseof easeof interpretation and will be the
for
future
one used
analyses. The extraction of the principle components produced four
factors with eigenvalues greater than one, explaining 42.5 % of the original variance.
Approximate percentage of variance for each of the factors is as follows:
12.1%, Factor 3 10.2% and factor 4 7.8%.
Factor 1 12.8 Oo, Factor 2
Cronbach's alpha on the fourth factor indicated that by dropping the negativek
loaded denial the internal consistency of the factor would greatly improve. Therefore,
denial was taken out of the factor leaving the factor to be comprised of positive maternal
representation and affection.
In order to later apply the factors for a story by story analysis, each factor extracted
by the varimax rotation was calculated per story and the means, standard deviations and
3.18.
below
Tables
3.17
and
in
alpha scores are presented
Table 3.17 : Meansl standard deviation, range and Cronbach's alpha per story by factors
(n=96)(nz--RQ)
two
qnd
nne
Mean
(SD)
Spilled Juice
1.9
(. 34)
Mother's Headache
1.6
(. 3)1)
Three's a Crowd
1.7
(. 38)
Burnt Hand
1.8
(. 34)
Lost Keys
1.8
Seet Shop
1.6
(. 29)
Scparation.
1.6
(. 39)
Reunion
1.8
(. 40)
Bathroom Shelf
1.8
(. 30)
Exclusion
_
Biscuit Till
1.7
(. 40)
1.8
_
19)
,Votal
disci pline /p unishment

(SID)
Range
Mean
F2
F1 quality/op en response
Story
1.8
Range
67 -2.333
.
67 - 2.33
.
67-2.50
.
1.0-2.33
50-2.33
.
50 - 2.33)
.
50-2.28
.
50 -2.28
.
50 -2.33
.
3)
.50-2.33
50 - 2.33
.
59- 116
cx
46
.
55
.
40
.
40
.
52
.
40
.
34
.
3'
.6
60
.
50
.
56
.
89
.
28
.
12
.
11
.
23
.
09
.
.30
02
.
04
.
18
.
10
.
n1
-)
1.17
(. 29)
0- 1.0
(. 19)
0-. 67
(. 17)
0-. 83
(. 25)
0- 83
.
0-. 83
(. 18)
(. 2 )
-3)
(. 08)
0-. 83'
0-. 50
(. 24)
0- 83
.
0-. 67
(. 19)
0-. 67
(. 26)
0 - 1.0
(. 10)
02-.45
.
(. 15)
Cx
75
.
65
.
59
.
72
.
74
.
58
.
55
.
82
.
74
.
74
.
63
.
93
Table 3.18 : Means, standard deviation, range and Cronbach's alpha per story by factors
three (n=89) and four (n=86).
Story
F3
negative /controlling
Mean
(SD)
Range
Spilled Juice
48
.
49
.
53
.
Mother's Headache
Three's a Crowd
Sweet Shop
Separation
Reunion
Bathroom Shelf
Exclusion
Biscuit Tin
Total
(. 23)
(. 28)
(. 30)
Burnt Hand
Lost Keys
(. 26)
53
.
59
.
57
.
57
.
58
.
57
.
54
.
1,54
(. 30)
(. 28)
(. 26)
(. 24)
(. 23)
(. 21)
(. 25)
(.20)
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
20.
1.33
1.40
1.50
1.60
1.40
1.40
1.27
1.40
1.40
1.20
1.25
20- 1.23
.
F4 p ositive maternal rep resentation

Mean
(SD)
Range
35
.
38
.
51
.
47
.
54
.
50
.
32
.
19
.
26
.
19
.
20
.
92
.
04
.
04
.
08
.
23
.
15
.
08
.
05
.
24
.
25
.
10
.
06
.
1.12
(. 16)
0-1.0
(. 15)
0-1.0
(. 21)
0-1.0
(. 29)
0-1.0
(. 29)
0-1.0
(" 1)
0-1.0
(. 17)
0-1.0
(. 40)
0-1.0
(. 33)
0-1.0
(. 25)
0-1.0
(. 20)
0-1.0
(. 12)
0-. 50
79
.
47
.
39
.
-),s
.
47
.
3,7
6
-3)
85
.
40
.
33
.6
48
.
factor
Response
Quality/Open
the
from
the
that
be
and
tables
It can
these
seen
40)
high
(-,
>
have
factor
alphas when
the most consistently
.
calculatedacrossstories.
Intercorrelations betweenstories
factors
four
between
intercorrelations
the
below
the
the
of
each
The tables
shov,high
likelihood
that
the
'Hicy
score
on
a
child
will
about
information
)ive
(
cleveii stories.
he
high
the
that
on
same
measure
in
or
she
scores
given
I-actor
story
one
in
a ccrtain
another storv.
94
Table 3.19
intercorrelations
between the eleven stories for the first
the
shows
factor quality/open response (composed of the variables role of parent, responsivity to

examiner, narrative coherence, joy, direct vs indirect performance stlyle and the child's
understanding of the conflict). The correlations are relatively high ranging from 23 - 69.
.
.
Table 3.19 : Intercorrelations between the eleven story stems for factor 1, quality/open
resvonse to task.
FI SJ
FI MH
FI 3C
FI
Fi LK
FI SS
Fi SN
F] RN
FI BS
FI EX
FI BT
BH
FI SJ
49
.
FI MH
FI 3C
31
.
39
.
42
.
31
.
43
.
44
.
35
.
33
.
F] BH
FI LK
FI SS
FI SN
FI RN
FI BS
FI EX
1
FI BT
50
.
36
.
49
.
29
.
36
.
34
.
48
..
28
.
28
.
28
.
42
.
34
.
30
.
23
.
28
.
23
.
28
l*
1
43
.
45
.
32
.
42
.
53
.
38
.
38
l*
58
.
44
.
53
.
12
57
.
35
.
4
48
.
47
..
56
.
41
.
46
.
58
.
69
.
55
.
1.58
68
.
61
.
1.63
65
.
1.67
l*
58
11
The intercorrelations of the eleven stories which involve the second factor, that of
discipline
and punishment
(composed of the variables excludes other, physical
disciplining
mother
punishment, shame other, verbal punishment and representations of a
for
be
lower
first
They
in
3.20.
Table
below
than
the
tend
to
father)
much
are shown
and
factor. The range is from -.02 - .29 with one exception. The correlation between story 7,
high
61.
is
8,
at
quite
reunion
and
separation,
.
95
Table 3.20: Presents the intercorrelations between the eleven story stems for factor
discipline/punishment.
F2 SJ
F2 MH
F2 3C
F2 BH
F2 LK
F2 SS
F2 SN
F2 RN
F2 BS
F2 EX
00
.
00
.
16
.
06
.
18
.
22
.
F-' BT
F2 SJ
F2 MH
F2 3C
29
.
12
F2 BH
11
F2 LK
18
F2 SS
15
23
.
13
.
06
..
lo
F2 SN
15
12
F2 RN
F2 BS
F2 BT
.
25
.
21
.
06
.
27
.
.
.
.
.
.
18
13
17
15
05
-.
13
.
08
.
05
.
15
.
22
.
04
.
17
.
16
.
29
.
09
.
12
.
os
12
.
20
.
07
-.
22
.
13
.
02
.
14
-.
08
.
09
-.
07
1.
05
-.
09
.
20
.
61
.
20
.
02
-.
.
16
Table 3.21 presentsthe intercorrelations of the third factor (composed of variables

investment
the
in
performance, control and
aggression,
physical
response,
negative atypical
from
33
high,
be
70.
factor
for
The
to
lack
tends
this
quite
range
of anxiety).
a
. -.
for
factor
between
intercorrelations
three,
the
Presents
Table 3.2 1:
the
eleven story stems
/control
neuntl ve.
SJ
F-33
_
F') MH
F-')
-)'C
_
F3 BH
f33 LK
F') SS
f3 SN
61
.
45
.
70
.
50
.
.-16
47
.
ll
Al
A1
F3 SN
F3 3C
F3 BH
F)' LK
F3 SS
68
.
64
1.
50
.
I
.
48
.
62
.
59
.
49
.
47
.
66
.
67
.
54
.
65
..
51
.
60
.
49
.
41
.
50
.
5-1
.
56
.
70
.
51
.
I
.42
.
JI
1
64
.
58
.
68
.
50
.
59
.
SJ F3 MH
F-33
_
F'3 RN
F') BS
F3 F-\
_
F') BT
I ino-
.-12
A5
.63
.45
A-i
1
.39
433
.
.55
57
.
F3 RN
F3 BS
F3 EX
F3 BT
63
.
.-48
62
.
55
.
54
.
50
.
:l
96
Factor four, Table 3.22 below, composed of the variables affection and positive
maternal representation, correlates at a relatively low level with a range of -.02 - 55.
.
Table 3.22:
Presentsthe intercorrelations between the eleven story stems for factor four,
positive maternal representation.
F4 SJ
F4 MH
F4 3C
F4 BH
F4 LK
02
.
12
.
oo
-.
09
.
04
-.
F4 SS
F4 SN
F4 RN
F4 BS
28
.
07
.
21
.
09
-.
1
30
.
03
-.
08
-.
F4 EX
F-4BT
07
1-.
F4 SJ
F4 MH
05
-.
F4 3C
08
.
19
.
04
-.
F4 BH
F4 LK
F4 SN
49
.
06
-.
F4 RN
17
F4 SS
F4 BS
lo
F4 EX
08
-.
F4 BT
10
09
.
22
.
15
.
09
-.
07
-.
1
--14
06
.
30
.
07
-.
1
17
18
.
04
.
12
.
44
.
25
.
01
.
1.01
12
07
-.
29
.
1.00
22
.
io
12
.
55
1.
12
lo
28
.
04
-.
21
.
1
05
02
-.
1
The intercorrelations of the stories for the factors described as 'quality/open

factors
described
for
higher
the
those
than
to
tend
I
of
response and 'negative/controlling'
first
involve
The
ipline/punishment.
two
the
'disc
and
'positive
representation'
maternal
as
involve
data
the
themes
two
the
content
and parental
mostly
second
whereas
pertormance
be
tends
to
It
that
similar across all
a child's performance style
representations. is possible
tell
the
type
themes
the
more
about
representations
us
and
of story
while
the story stems
For
-positive
this
example,
a
child's
score
on
narrative.
maternal
to
elicit
Used
Stern
repivscntation'
does not tend to be highly correlated with their 'Positive maternal
97
representation' score on any other story. This might suggest that the themes of affection
and positive maternal representations are more related to a specific story steiii than
indicative of the internal world of the child who told the narrative.
Creation of Scalesftom thefactor scores
Scales from the factor scoreswere then createdby aggregatingthe variables which
loaded with absolute values greater than 0.50 and retaining the sign of the loading. Table
3.23 presents the variable composites of each of the scales and their Cronbach's alphas
summed across the eleven stories.
Table 3.23 : The composition of the four factor scales

Factor I
Quality/open
Factor 2
Discipline and
punishment
disciplining
to
mother responsivity
+
examiner
+
verbal punishment
direct vs indirect
+
father
disciplining
performance style
+
+
excludes other
child's understanding
+
of conflict +
shame other
+
coherence
narrative
+
physical punishment
response
role 2tLpare
:Lnt
76
.
74
.
Factor 3
Controlling and
Factor 4
Positive maternal
negative
representation
investment in
performance
+
physical aggression
+
control
+
negative atypical
response
affection
+
positive maternal
representation
anxiety
58
.
98
Lack of anxiety may seem strange at first until one considers that the children*s
apparent lack of anxiety may be being enacted in the controlling and aggressivebehaviour
Z,
and responses.
In conclusion,, the data reduction section demonstrates that a set of reliable and
be
consistent scales could
produced which could, in turn, be aggregatedacross all stories
be
basis
factor
themselves
that
these
the
then.
could
aggregated
on
of
scales
and
analysis
)
to form interpretable constructs with high internal consistency. Hence, the psychometric
in
incomplete
previous research were addressed.
properties
The subsequent chapters will employ the four new factor scales in relation to the
Stem
Battery.
Story
MacArthur
the
children's responseson
99
CHAPTER FOUR
THE EFFECTS OF DEMOGRAPHIC

CHILDREN'S
FACTORS ON THE
RESPONSES TO A NARRATIVE
TASK
4.1 INTRODUCTION
k k (n
Chapter Three responded to a need for more detailed psychometric information

regarding the MacArthur Story Stem Battery and the accompanying coding system, the
MacArthur Narrative Coding Scheme. While the MacArthur Stor,Stem Battery is a most
for
investigating
promising method
aspects of the young child's internal world it is also a
complex method and many elements need to be teasedout in order to make senseof it. In
the literature describing the few studies with story stems there has been a call for much
information
descriptive
data
different
Klitzing,
KelsaN,
(von
regarding
needed
within
samples
& Emde, 1998). In an attempt to provide some of this much needed infori-nation and
begins
by
task,
this
the
the
stem
chapter
exploration
examining the
about
story
clarification
factors
influencing
demography
language
as age,
of
abilities and
possible effects of such
battery.
Parent's
to
the
the
and children's age,
story stem
gender upon
level
language
of education, social
abilities and parent's
parent's and children's expressive
be
temperament
considered.
will
class and child's
1.1 Age
Children as young as three years of age have been able to be cooperative and
have
to
task
the
to
produced
meaningful
responses
and
stem
story
productive in responding
dilemmas
for
to
moral
and affectively
empathic, prosocial responses
the stories that probed
By
boys
&
Emde,
1990).
the
(Buchsbaum
early
school
years
and
girls
scenarios
charged
for
basic,
describing
internal
have
a
common
vocabulary
acquired
states
classes
social
across
and evidence of the use of multiple narrative structures has been found in the spontaneous
stories of four-year old children (Tarullo, 1994; Hudson, Gebelt, Haviland & Benti'v(pa.
1992). When looking at five-year-old children's narratives it has been found that they are
differentiated
more
and detailed in their depictions of family members than three-year-olds
(Bretherton, Prentiss & Ridgeway, 1990). As the children this sample are five years of a(-,c
have
and, consequently,
more experience with language and narratives than three-and-four
is
year-olds, it expected that they will be able to produce coherent and meaningful stories
dilemmas
the
about
and conflicts presented to them in the story stems. Although the
differences in the ages of the children participating in this study is small, development
fast
in
rate young children, and it is possible that the older children will produce
moves at a
higher
There
than
that
the
are
of
a
quality
and
open
more
younger
children.
is no
narratives
prediction made regarding the effect of parental age upon the outcome of the narrative task.
4.1.2 Language
With regard to language, several studies which have quantitatively studied the
MacArthur Story Stem Battery have controlled for children's language use because it is
influence
language
type
the
of responses given
that
ability
will
children's
possible
&
Nir,
Warren
Emde,
1997;
Warren,
Oppenheim,
1996;
Wambolt,
&
Emde
(Oppenheim,
Oppenheim & Emde, 1996). One of the studies looking at children's representationsof their
between
language
found
correlation
expressive
the
and
one moderate
story stems
mothers' in
5
(Oppenheim,
Emde
&
Warren,,
1997).
at
age
years
mother
positive representations of
bevvveen
language
(as
by
foLind
the
expressive
\nother
cot-relations
measured
moderate
ca
few
Test)
Vocabulary
One-word
Picture
and a
story stem variables such as
expressive
for
directness
but
le.
themes
of
st\
no
correlations
such
as
coherence and aggression
in
(Oppenheim,
themes
the
to
and relatedness
examiner
investment performance, prosocial
Nir, Warren & Emde, 1997). The current study also makes use of an expressive language
it
is
that
particularly the children's expressive
assessment which seems appropriate given
language skills that are being utilized in this task. The potential effects of both the children's
language
before
birth
(assessed
the
the
ability
of the target child)
parent's expressive
and
language
is
high
investigated.
be
It
that
the
on
expressive
predicted
children scoring
will
higher
that
quality and are more open than children
are of a
assessmentwill produce stories
be
formal
language
No
low
the
made with regard
predictions will
assessment.
on
who score
to the other three factor scales.
4.1.3 Infant Temperament

The main categories classifying child temperament have been derived from
i
development,
infant's
ing
their
their
resulting
in
II
regardi
concerns
about
parents
interviewing
difficult
handling,
to
to
dimensions;
slow
warm up
and
manage
these
three
are ease of
(Thomas & Chess, 1977). The last of these, slow to warm up, is a measureof inhibition and
first
late
into
the
be
found
has been
childhood when classified after
to
especially persistent
1984).
(Kagan,
In
be
it
a
birthday, although signs of can
observed at about eight months
five
to
the
task
similar
ones
stories
using
completion
three-year-olds
story
and
a
study of
by
(as
between
the
in
scales
measured
and
shyness
sociability
this
study, correlations
used
filled
found-,
by
18
Inventory)
Temperament
months were
out
mother at
Colorado Child
03
(Bretherton,
to
the
negatively
story scores
sociability relating positively and shynessrelating
Ridgeway & Cassidy, 1990). It is therefore, predicted that those children scoring,higher on
the temperament measure, meaning that they are more difficult/inhibited,
have
lw,
will
-er
less
judged
difficult/inhibited
than
and
open
responses
children
not
quality stories
by their
parents.
4.1.4 Social class

Reports of studies using the story stemshave mainly usedpopulations of middle-class
families and have not reported on the associations of the children's responseswith social
be
how
difficult
It
to
social class might
related to the outcome of the
predict
is
class.
literature
however,
pertaining to internalization of parental values might
children's stories,
be relevant to this question. For example, lower socioeconomic statusparenting is associated
less
lower
discipline,
verbal
warmth and affection and
with rigidity, power-assertive styles of
being
This
1997).
that
(MacDonald,
the
possible
children
case,
it
is
interaction with children
have
lower
from
tend
themes
to
the
of
more
might
group
class
social
whose parent's are
discipline and punishment in their stories and fewer themes of positive maternal
from
in
this
However,
that
the
study are overwhelmingly
subjects
given
representations.
to
the
effect
are
made
with
regard
predictions
no
classes,
socioeconomic
middle
and
upper
of social class on the narrative task.
4.1.5 Gender
Sonic gender differences, in aggressionfor example, are well attestedcross-culturally.
Z--
C)+
few
Vocabulary
Test)
One-word
Picture
and a
expressive
story stem variables such as
but
for
directness
themes
correlations
no
coherence and aggression
such as
of style.
in
investment performance, prosocial themes and relatedness to the examiner (Oppenheim,
Nir, Warren & Emde, 1997). The current study also makes use of an expressive language
it
that
assessmentwhich seems appropriate given
is particularly the children's expressi%,
c
language skills that are being utilized in this task. The potential effects of both the children's
language
before
birth
(assessed
the
the
expressive
parent's
ability
of the target child)
and
is
investigated.
It
high
language
be
that
the
predicted
children
scoring
on
expressive
will
higher
I
dren
than
that
quality and are more open
chi
are of a
assessmentwill produce stories
formal
be
language
No
low
the
predictions will
made with regard
on
assessment.
who score
factor
three
the
to
scales.
other
4.1.3 Infant Temperament

derived
from
have
been
temperament
The main categories classifying child
in
development,
infant's
their
their
resulting
regarding
concerns
about
parents
interviewing
difficult
handling,
to
to
dimensions;
and
slow
warm up
manage
these
of
ease
three
are
inhibition
is
last
The
1977).
to
Chess,
&
these,
and
(Thomas
slow wann up, a measureof
of
has been found to be especially persistent into late childhood when classified after the first
birthday, although signs of it can be observed at about eight months (Kagan, 1984). In a
five
task
to
the
stories
similar
ones
completion
using
a
story
and
three-year-olds
of
study
between
(as
by
the
sociability
and
shyness
scales
measured
correlations
this
stud,.,,,
used in
filled
by
found.
Inventory)
18
Temperament
Child
out
mother at
months were
Colorado
C)s
sociability relating positively and shynessrelating negatively to the story scores (Bretherton.
Ridgeway & Cassidy, 1990). It is therefore. predicted that those children scoring higher on
the temperament measure, meaning that they are more difficult/inhibited.
have
lo\\eiwill
quality stories and less open responsesthan children not judged difficult/inhibited
by their
parents.
4.1.4 Social class

Reports of studies using the story stemshave mainly usedpopulations of middle-class
tamilies and have not reported on the associations of the children's responseswith social
difficult
how
be
It
to
predict
social
is
class
might
class.
related to the outcome of the
literature
internalization
however,
to
pertaining
of Parental values might
children's stories,
be relevant to this question. For example, lower socioeconomic status parenting is associated
discipline,
lower
less
warmth and affection and
verbal
with rigidity, power-assertive styles of
being
it
This
1997).
(MacDonald,
the
case, is possible that children
interaction with children
have
lower
from
tend
to
themes
the
group
might
social
class
more
of
are
parent's
whose
discipline and punishment in their stories and fewer themes of positive maternal
in
from
However,
this
that
the
study are overwhelmingly
subjects
given
representations.
to
the
are
made
with
predictions
regard
effect
classes,
no
socioeconomic
Lipperand middle
of social class on the narrative task.
4.1.5 Gender
Some oendcr differences, in aggressionfor example, are well attestedcross-culturally.
\o (o
Those researchersstudying gender differences agreethat boys and girls seem to be speaking
different
different
dimensions
have
the
these
voices about
ways of experiencing
in
world yet
in
difficult
to
capture
experimental situations and rarel-,,,translate into significant
proven
differences in groups means (Tarullo, 1994). For most socloaffectl\'e measures, boys and
look
different
being
than
more
alike
on
will
average,
girls
aggression
a notable exception
(Tarullo, 1994). In the only study to report on gender differences and the story stems. some
differences were found between boys and girls (Oppenheim, Nir, Warren & Emde, 1997).
At age four-and-a-half girls were identified as having more prosocial themes and fewer
found
five.
both
boys
but
differences
At
41/2
these
than
themes
at age
age
were not
aggressive
boys
direct
but
higher
had
5
1/2
than
to
the
and
a
more
style
relatedness
examiner
girls
and
there were no differences found between gendersfor investment in performance or on their
language
abilities.
expressive
As a result of the findings from the above studies, it is
four
five-year-olds
found
be
differences
the
this
on
sample of
in
will
predicted that no gender
in
literature
been
has
little
the
factors.
Because
about
reported
so
information
stem
story
be
individual
hoc
differences
the
variables will
analysis of
a post
story stems and gender
boys'
it
is
that
As
the exception,
narratives will contain more instances
expected
conducted.
for
direction
in
is
There
the
remaining
the
either
than
expectation
no
girls.
of aggression
themes coded for.
4.2 METHOD
4.2.1 Design
during
last
demographic
the
trimester
the
The parent's
variables were collected
of
Z--
target children's pregnancy in Phase I of the London Parent Child Project. The original
sample consisted of a group of 100 cohabiting parents who were recruited prior to the birth
first
their
of
child. The parent's expressive language ability ,,vas assessedduring the first
prenatal visit using the Short Form of the Mill Hill Vocabulary Scales, designed and
by
Raven
(1986)
(see Appendix B. 1). The MacArthur Story Stem Batterv
validated
et al
(Appendix A. 1) and the Bus Story test (Renfrew, 1991) (Appendix B. '3) was administered
to the five year old children during the first visit of Phase III of the London Parent-Child
Project. The MSSB has been described in detail in Chapter Three.
4.2.2 Subjects
Mothers
The original sample (n = 100) of mothers turned out to be mainly an older group of
first-time mothers, well-educated, white and predominantly middle-class; 70 of the women
held university degreesand all 100 had attended secondaryschool at the time of recruitment.
The expectant mothers' median age was 31 years (range of 22 - 42). Seventy-five of the
bom
Scotland
Ireland
15
UK
in
10
bom
England,
the
or
while
were
outside
in
women were
from
Western
Anglo-Saxon;
British
latter
This
Ireland.
the
and
six
was
primarily
group
or
Commonwealth countries of Australia, New Zealand and Canada, two from the United
States,one from South Africa, four European with only the remaining two from non-Westem
father
Eighty-two
the
to
the
expectant
woman were married
at the time of
of
cultures.
At
length
the
the
prenatal assessment,
median
of
recrUltnient or rnarried subsequently.
father
5
(range
1-19).
Twenty-one
the
together
expectant
was
years
with
of the
residence
I
65
(professional
11
class
and
social
managerial);
social
class
woman represented
ol
(intermediate occupations): 9 social class III (skilled occupations). and 5 social class IV
(partly skilled occupations) according to the criteria of the UK government Office of
Population Censuses and Surveys (1980). Sixty-fi,,,,e of the women were in the middleincome group, 21 in the upper income group and 14 in the lower income group.
Fathers
Like their female partners, the sample of expectant fathers (n=100) constituted a
highly homogeneous educated, middle and upper-middle class group. Seventy-one were
had
further
educated,
only
university
one
no
education at the time of recruitment. Their
33
20-57).
(range
Eighty-seven
median age was
of the expectant fathers were from England,
tive were from Scotland or Ireland and eight were born outside the UK, though none of these
different
Twenty-five
being
cultures.
of the men representedsocial class
markedly
countries
I (professional and managerial), 57 social class 11(intermediate occupations); 16 social class
III (skilled occupations); and 2 social class IV (partly skilled occupations), according to the
Censuses
Surveys
(1980).
In
Office
Population
Government
UK
terms
the
and
of
criteria of
for
60
the
the
the
to
of
men
were
in
middle
mothers,
the
above
referred
groups
income
of
income
lower
(see
16
income
24
the
the
group
also
and
in
group
upper
in
income group,
Steele, H., 1991, Steele, M., 1990; Fonagy, Steele & Steele, 1991).
Chilch-en
Ninety-two children returned with their parents for the third phaseof the Project and
battery
Bus
Story
Forty-three
the
test.
the
story stem
and
of
eighty-nine actually completed
forty-nine
boys.
The
61
the
and
mean
age
of
children was
the sample children are girls
\)Ij
months (range 59 - 65 months). The mean ages of the boys and girls did not differ
signi icantly (girls mean age 61.3 months; boys mean age 61.1 months).
4.2.3 Procedure
As stated,,the parent's expressive languageability was assessedusing the Short Form
Mill
Hill
Vocabulary
Scales.
The variable scores the number of words both correctly
the
of
defined and used appropriately in sentences,yielding a single continuous score of verbal skill
(Appendix B. 1).
The children's temperament was assessed at twelve months using the Infant
Characteristics Questionnaire (Bates, Freland & Lounsbury, 1979; seeAppendix B. 2). This
infant
difficult
be
how
their
they
to
to
perceive
relative to the
instrument asks parents rate
i
has
been
instrument
first
The
throughout
the
this
well-establ shed
validity of
averagechild.
independently
fathers
Both
life.
the
completed
questionnaire which
and
mothers
of
year
their
the
to
parents'
perceptions
of
of
child's
assessment
a
global
yield
summed
'difficultness'.
The Bus Story test (Appendix B. 3) was presented to the children following the
MacArthur
Story
Stem
Situation
Battery.
Strange
Modified
the
the
and
administration of
language
(Bishop
is
Story
Bus
The
test a measureof performance on standardized
responses
involved
It
1987).
the
Edmundson,
child
a
series
showing
of colour sketchesportra,,,Iing
and
befall
bus.
When
finished
'bus
that
tester
the
telling
the
a
calamities
the adveiitures and
lo
invited
the
to retell the story -,ith resort to only the pictures. The child's
story'
child was
for
narrative response was audio-recorded and transcribed -,,,
erbatim
subsequent scoring in
terms of three dimensions: 1) averagesentencelength, 2) averagenumber of subclausesused
3)
These
three scaleswere then summed to provide
per sentenceand quality of information.
Story
Bus
language
development
both
aggregate,
reflecting
the
a
robust
measure
of
a
early
in
for
The
the
test
and
particularly,
expressive
receptive,
modes.
preschool age children has
been in continuous over the last twenty years in the United Kingdom and there are \,al1d1tY
be
language
delays
in
the
test
to
uniquely successful in predicting
results showing
school age
(Renfrew,
1991).
children
4.3 RESULTS
The results are described in two main sections.The first section considersthe parent's
four
Next
the
the
to
the
considered is
story stems.
responsesof
and children's age in relation
language
to
dimensions
factors or
of the children's story stem responsesas possibly related
level
to
the
the
both
the
parents
prior
of
just
educational
and
parents and children
abilities of
birth of the child. This is followed by descriptions of the correlations of infant temperament
first
in
Finally
five
this
task.
the
section,
are
narrative
on
age
at
responses
with children's
factors.
four
to
the
story stem
the results of parent's social class in relation
The second
their
to
the
detail
deals
responses
the
gender
upon
child's
of
possibleeffects
with
in
section
narrative task.
4.3.1
Age
Table 4.1 below shows the correlation coefficients and the levels of slunificance of
father's
four
factors
the
age with
of the children's responses
child's age, mother's age and
to the story stems.
Table 4.1 : Bivariate correlations and levels of significance of mother's, father's and
four
factors.
the
story stem
children's a2e and
Correlation
Coefficients
(p)
mother's age
father's age
Quality/open response
004
(.
97)
-.
(.
143
18)
-.
087
.
(. 41)
89
Discipline/punishment
(.
023
83)
-.
014
(.
89)
-.
128
(. 23)
86
Control ling/negative
135 (. 21)
Note :r
age
(. 81)
165
-.
(. 12)
89
000 (. 99)
050
(. 64)
86
025
.
035 (. 75)
.
children's
levels
Pearson
the
in
significance
correlations with
values are
based
tests
two-tailed
on
are
;p
values
parentheses
between
4.1
Table
either
above indicates that there are no significant correlations
MacArthur
factors
four
to
the
the
the
responses
children's
of
and
age
or
parent's
age
child's
Story Stems.
4.3.2 Expressive language and level of education

In order to explore the possibility of expressive language abilities of the children and
these
the
variables was undertaken.
of
analysis
a
correlational
results
their parent's affecting
investigate
the
to
possibility of an nfluence
conducted
IIIII
was
is
analys'
SimilarIv, a correlational
I
of the parents' level of education.
The data for the children's expressive language, as measured by the Bus Story test,
interval
data
is
and so the Pearson correlation was used. Table 4.2 below sho,,,
-s the
correlation coefficients and levels of significance of the aggregate Bus Story score as %N,
cll
as the three sub-scales of sentence length, number of clauses and the quality of the
information provided with the four factors derived from the MacArthur Narrati,,,,e Coding
Scheme (as described in Chapter Three).
Table 4.2 : Bivariate correlations and levels of significance for the Bus Story scales and the
four story stem factors.
Correlation
Bus Story
108 (.31)
Sentence length Clauses

134 (. 22)
.
164 (. 13) 000 (.99)
.
.
Controlling/negative
039
003
97)
(.
71)
(.
-.
.
Positive maternal representation 1 113 (. 29) 1 161 (. 14)
.
.
Note :r
Coefficients
108 (. 33)
124 (.26)
(p*)
Quality of info
153 (. 16)
89
196 (. 07)
86
.
.
049 (. 65) -.054 (. 62)
.
164
(.
14)
082
(.
46)
11.
levels
in
Pearson
the
significance
are
correlations
with
values
parentheses;
based
on two-tailed tests
p values are
It can be seen from Table 4.2 that there are no significant associations between the
Story
Bus
language
four
factors
the
the
test
the
on
any
of
scales
of
abilities
and
children's
deriN,ed from the MacArthur Narrative Coding System of the story stems. There is one trend
factor
between
the
and one of the sub-scales, quality of
reported
infori-nation. However, as it is a trend and the aggregate Bus Story score
not significant,
89
86
the effects of language will not be controlled for in any further analyses of the data.
The data for the parent's expressive languageand educational level is ranked data and
Spearman's
rho was used to calculate the correlations of parent's expressive languageand
so
four
level
factors.
Table 4.3 below shows the correlation coefficients
the
with
educational
for
levels
the
these calculations.
of significance
and
Table 4.3 : Correlations and levels of significance for parents' expressive language and
four
factors.
level
the
with
storv stem
education
Correlation
Coefficients (p)
Education level
Expressive language
father
mother
165 (. 12)
072
(.
50)
-.
081 (. 44)
.
16)
152(.
-.
(.
87)
017
-.
99)
000(.
-.
009(. 93)
.
mother
Positive maternal
representation
Note :r
father
156 (. 14)
89
006(. 95)
.
071
(.
51)
-.
86
074
(.
48)
-.
154(.
14)
-.
055
(.
60)
-.
89
(.
18)
142
-.
67)
046(.
-.
(.
040
71)
-.
86
levels
Speannan's
the
significance
in
rho correlations with
values are
based
tests
two-tailed
on
are
p
values
parentheses,
As can be seen in Table 4.3, there are no significant correlations.
4.3.3 Infant
temperament
father
by
using the
As stated, infant temperament was rated separately
mother and
k\L
Infant CharacteristicsQuestionnaire.Table 4.4 below showsthe Pearsoncorrelationsof the
four story stem factors with both mother*s and father's ratings of the child's temperament
at one year.
Table 4.4 : Correlations and levels of significance of infant temperament and the story stem
factors.
Correlation
Coefficients
(p)
mother's rating
father's rating
(.
093
38)
-.
034
(.
75)
-.
89
Disc ipl ine/pun ishment
041 (. 70)
.
195 (. 07)
86
068 (. 52)
.
064 (. 54)
.
89
024 (. 82)
.
074 (. 49)
-
86
Note :r
levels
Pearson
the
significance
values are
correlations with
in parentheses;
based
on two-tailed tests
p values are
Again, there are no significant correlations for either mother's or father's rating of
factors.
the
temperament
with any of
infant
There is one trend noted between the
Disc i pl ine/Punishment factor and father's rating of infant temperament, however, they are
future
for
be
analyses.
in
controlled
not significant and will not
4.3.4 Parent's social class

The possible effects of the parent's social class on the children's responsesto the
is
data
Because
the
an
measured
on
interval
class
social
next.
considered
was
stems
story
homogeneity
distributed
there
of
variance
a
one-way
and
is
the
eveifly
are
scores
scale,
Group
Because
III,
those
th's
to
I
question.
examlne
of
conducted
was
variance
ot
analysis
\\s
skilled occupations, and Group IV. those of partly skilled occupations, contained too tc\\
I
members (8 and 3 respectively) to make the statistical analysis reliable in both the mother
father
and
samples these two groups were combined into one group (Group III beloNA-)
containing those whose occupations are skilled and partly skilled.
Levene's test for
homogeneity of variances was observed and seen not to be significant.
Table 4.5 below considers the possible effects of the mother's social class on the
children's responsesto the story stems and shows the means. standard deviations, degrees
freedom
f-ratio
the
and
of
and f-probabilities of the factor scoresgrouped by mother's social
class.
Table 4.5 : Means and SDS of factor scores grouped by three classifications of mother's
social class.
Means (SD)
Factor
df
11
111
1.8596 (. 15)
1.7507 (. 28)
1.7502 (. 19)
Disc ip II ne/pun ishment
1652(. 08)
1641 (. 10)
Control I Ing/negative
5043 (. 16)
.
1386(. 14)
Positive maternal
11representation
f-ratio
f-prob
2,84 1 1.41
24
.
1804(. 10)
2,81
12
87
.
5363 (. 18)
.
6248(. 29)
.
2,84
1.38
25
.
1089(. 10)
1694(. 12)
2,81
1.38
25
.
Note: p values are based on two-tailed tests
Table 4.5 indicates that there is no significant effect of mother's social class on the
hoc
in
Scheff6's
factors.
four
test
was not observed this case as there are
post
either of the
no sigonificant main effects.
k\(e:
for
father's
A one-way analysis of variance -.,,,
the
as
also
conducted
,,
social class to
consider the possible effects of the paternal social class on the children's responsesto the
below
4.6
Table
shows the means, standarddeviations, degrees of freedom and
story stems.
the f-ratio and f-probabilities of the four factor scores grouped by father's social class.
Table 4.6 :Means and SDSs between factor scoresgrouped by three classifications of father's
social class
Means (SD)
Factor
Positive maternal
1representation
Note: *p : 05
.
f-prob
11
111
1.8581 (. 16)
1.7603 (. 24)
1.7096 (. 36)
2,84
1.80
1481 (. 08)
1576(. 09)
2212(. 11)
.
2,81
11
-3).
04*
.
4875 (. 15)
.
5452(. 18)
.
5968(. 26)
.
2,84
1.49
23
.
1612(. 14)
1054(. 10)
1273 (. 11)
2,81
1.66
19
.
II
f-ratio
df
.
I
.
I
17
(p values are based on two-tailed tests two-tailed)
Table 4.6 above indicates that, unlike mother's social class, there is an effect of
father's
to
father's social class on the children's responses.
is related
hoc
Scheff6's
that
04.
test
level
confirmed
no
and
observed
was
post
of
p=
a
social class at
.
father's
level.
05
As
this
different
the
analysis
a result of
p: .
at
two groups are significantly
future
for
the
be
involving
analyses
in
controlled
will
social class
actor.
4.3.5 Gender
for
to
this
most socio-affective measures
chapter,
As mentioned in the introduction
boYsand girls will look more alike than different on average.

aggression being the notable
exception. This section will examine the issue of gender differences in some detail and will
look at the possibility of an association between gender
and the children's responseson the
narrative task.
To first explore the question of whether there were differences between the bov's
and
I
language
girl's expressive
ability which might affect the type or quality of their narrative
responses,independent sample t-tests were calculated. Levene's test for the equality of the
variance was observed and the significance levels for equal groups was used as indicated.
The association between gender and each of the four scalesof the Bus Story was tested (see
Table 4.7 below).
Table 4: 7 : Means, T scores and levels of significance for the Bus Story scales grouped by
P-ender.
Means (SD)
Girls
df
t-value
Boys
Bus Story
40.32
(9.84)
39.89 (9.61)
98
22
.
Sentence length
11.23 (2.45)
10.64 (2.13)
82
1.17
24
.
(. 90)
82
1.01
31
.
27.93 (8.28)
82
16
87
.
Clause
1.02
Quality of info
(. 87)
28.233 (8.83)
1.22
82
Note: p values are based on two-tailed tests, standard deviations are in parentheses
The results obtained show that no significant differences were found between girls'
language
boys'
ability.
and
expressive
To explore the question of gender and the children's responseson the MacArthur
\\
Story Stem Battery independent sample t-tests were calculated to determine if the boy's
differed
from
the girl's on any of the four factors. Levene's test for the equality
responses
of the variance was observed and the signIficance levels for equal groups was used as
indicated. The association between gender and each of the four factors derived from the
factor analysis of the MacArthur Narrative Coding System was tested (see Table 4.8 belox\).
Significance levels reported are based on two tailed tests as there is no a priori hypothesis
differences
factor
to
the
regarding gender
in relation
scales.
Table 4.8 : Means, SDS and levels of significance by gender for the four storv stem factors.
Means (SD)
Discipline / punishment
Negative / controlling
df
t-value
86
girls
boys
1.8011 (. 21)
1.7546 (. 28)
87
1542 (. 08)
1746 (. 10)
84
.
5131 (. 19)
.
1443 (. 12)
-
.
5669 (. 19) 87
.
1067 (. 10) 184
-
.
98
.
1.30
1.48
39
.
33
.
19
.
14
.
Note: p values are based on two-tailed tests and standard deviations are in parentheses
The results obtained in Table 4.8 show that no significant differences were found
between gender and the four factors.
Becausevery little work has been published on the question of gender differences in
in
detail
by
it
decided
this
to
area
more
conducting
explore
relation to the story stems was
An
individual
to
hoc
the
gender.
additional
relation
variables
in
analysisexamining
a post
in
the
the
reliably
coded
were
not
included
that
were
which
of
variables
some
reason is
has
been
done
Some
factor
the
tactors as a result of
work
regarding content themes
analysis.
v)
and gender but none regarding the performance and affect themes (Woolgar. 1996). As the
literature about children's play themes in general has suggestedthat particular themes such
be
by
aggression
as
and prosocial responsesmight
affected
gender it is predicted that some
helping,
be
by
variables, such as aggression, empathy and
might
affected
gender.
Those variables which proved reliable were aggregatedacrossthe stories. Only those
internal
having
higher
that
than 40, ,,vere used in the
consistency,
variables with
is,
alphas
.
following analysis (as described in Chapter 3). Independent t-tests were conducted and the
levels
based
is
hypothesis
two-tailed
tests
there
reported
are
on
as
a
significance
no priori
Table
4.9
below
the
the
variables.
majority of
indicates the results of the
regarding
independent sample t-tests for each content and parental representation variable by gender.
The performance variables will follow.
Table 4.9 : Means, SDS and levels of significance for the content and parental representation
by
(n=86).
variables grouved
Pender
Variable
Means (SD)
Girls
df
t-value
Boys
1364
(. 15)
0949
.
(. 12)
84
1-39
Dishonesty
0977
.
(. 11)
1087
(. 12)
84
41
.
68
.
Guilt/reparation
0591
.
(. 09)
0316
.
(. 08)
84
1.45
14
0349
.
(. 08)
0838
.
(. 13)
75.78
2.12
1601
(. 17)
80.36
1.74
08
.
1798
(. 16)
79.91
4.05
000***
.
Affection
Excludes other
1045
(. 12)
Physical aggression
0614
.
(11)
0578
.
(. 10)
0672
.
(. 09)
84
44
.
66
.
Physical punishment
0568
.
(. 11)
1206
(. 14)
84
2.26
02
.
Disciplining father
0795
.
(. 07)
1364
(. 12)
74.02
2.55
01**
.
Disciplining mother
2955
.
(. 14)
2648
.
(. 18)
84
85
.
39
.
1186
(. 11)
84
1.233
Y)
.
.
.
.
Positive mother
1523
(. 13)
Excludes self
0909
.
(. 08)
0810
.
(. 10)
84
48
.
62J
.
Shame other
1136
(. 12)
1087
(. 14)
84
17
86
.
Verbal punishment
2750
.
(. 15)
2569
.
(. 18)
84
48
*
63
.
Note
*** p : 001
** p< 01
* P: 05
.
.
.
based
tests
two-tailed
on
are
p values
four
4.9),
(Table
from
there
be
table
and
parental
content
are
the
As can
above
seen
boys
between
the
the
differences
the
girls
across
and
that
signify
representationvariables
boys
in
the
of
responses
Unsurprisingly,
more
seen
was
aggression
physical
elevenstories.
Boys
11).
0614,
(mean
16)
also gave more
than girls
scl =.
= .
(mean = 1798, sd =.
.
0795,
(mean
12)
1364,
father
(mean
than
disciplining
=
girls
sd
=.
=
a
representations of
.
.
14)
1206,
(mean
and
sd
=.
=
punishment
07)
of
physical
instances
more
and enacted
sd=.
.
There
in
I')
0838,
thei
(mean
completions.
also
story
is
ir
sd
=.
=.
responses
negativc atvpical
17)
1,
160
than
the
theme
'excludes
to
other'
more
girls
of
boys
(mean
use
for
sd
=.
=.
a trend
ZN
Rk
(mean =. 1045, sd =. 12).
Table 4.10 below presentsthe results of the independent t-tests for the performance
by
gender.
variables grouped
Table 4.10 : Means, SDS and levels of significance for the performance variables by gender
n=89
Means (SD)
Variable
df
Girls
Boys
Role of parent
1.8337 (. 38)
1.7670 (. 50)
87
69
.
1.0665 (. 11)
1.0341 (. 18)
87
98
.
Directness of performance style
1.9290 (. 18)
1.8636 (. 31)
76.94
2.7184
(. 43)
2.7064 (. 47)
87
Involvement of examiner
1.4279 (. 27)
1.4545 (. 29)
2.2106
1.21
49
.
.23
12
90
.
87
44
.
66
.
(. 67)
87
13
89
2550 (. 16)
.
3030 (. 17)
.
87
1.34
Adaptiveness of response
1.8359 (. 42)
1.5966 (. 31)
87
3.02
Narrative coherence
2.7517
(. 55)
87
94
.
35
.
Denial
(. 70)
(. 48)
2.1913
2.6477
18
003
.
Control
0887 (. 17)
.
1282 (. 22)
87
91
.
36
.
Joy
5070 (. 26)
.
5088 (. 24)
.
87
03
.
97
.
Anger
0717
.
(. 06)
0436
.
(. 04)
87
2.32
Distress
0259
.
(. 04)
0234
.
(. 03)
87
29
.
77
.
Concern
0466
.
(. 07)
0423 (. 05)
.
87
30
.
76
.
Anxiety
6-245 (. 26)
5688 (. 29)
87
94
.
34
.
Note :*P! 05
.
** p ! 01
.
02*
.
(p values are based on two-tailed tests)
Table
4.10
looking
the
When
above shows that only two
performance variables,
at
Clk 01%
variables show significant differences between the boys and girls. Rather surprisingly, girls
expressed more anger in their stories (mean = 0717, sd =. 06) than boys (mean = 0436. sd
.
.
04).
They were also judged to have significantly higher adaptive responsesto the
=.
story
stem task (girls mean = 1.8359, sd =. 42; boys mean =1.5966, sd =. 3 1).
4.4 DISCUSSION
This chapter undertook to examine the effect of various demographic variables upoii
the children's responsesto the MacArthur Story Stem Battery task. There are several points
in
mentioned the results sectionof which to take note here.
The possibility that the older children would produce narratives of a higher and more
be
between
to
the
there
the age of the
case;
was no correlation
open quality proved not
The
task.
the
their
narrative
results showed that there was also no
responseson
children and
effect of either parents' age on the child's responsesnor was there an effect of either parents
is
in
fact
This
level
the
that the vast
the
result not surprising view of
on
results.
education
majority of parents are university educated.
As mentioned in the introduction, other studies of similarly aged. middle-class

language
between
found
the
abilities and the story
child's
significant associations
children
between
In
there
this sample of children
were no significant correlations
either the
sterns.
four
factors
language
derived
from
the
the
abilities
any
the
with
of
parent's
children's or
children's responses on the narrative task. It is possible that the lack of associations of
children's expressive language abilities to the factors is due to the fact that this studN,used
different
instrument
language
to
a
assess
ability and that it is not as sensitive to expressi've
language issues as those used in the studies mentioned. It is also possible that differences
in the children's scores are not large enough and could be explained by the
relative
homogeneity of this sample. Due to the non-significant result, subsequentanalysesof these
factors
for
story stem
will not control
child or parent language abIlIties.
Infant temperament was rated by the both parents independently of each other and
be
to
neither parents' rating proved
associated with the story stem factors. The above
findings tend evidence to support the conclusion that the story stems the children produce
factor
language
their
of
abilities, their age or their temperament.
are not a
There was one significant result regarding associations between the parents' social
factors.
had
Mothers'
the
social class
no associations to any of the
story stem
class and
factors, however, father's social class was related to one of the factors. Those fathers
belonging to the group of occupations defined as skilled or partly skilled were found to have
discipline
This
the
themes
to
and
punishment
in
narratives.
of
children who referred more
factor
but
for
for
the
the
to
not
made
predictions
corresponded
Positive Maternal Representation factor. The Discipline/Punishment scale includes less
harsh forms of discipline such as verbal punishment, however, it also includes harsher styles
The
the
the
other.
other
and
excluding
positive
result
Suchas physical punishment, shaming
,
lower
literature
by
that
be
the
socioeconomic status parenting is
mentioned,
explained
may
CID,
L
discipline.
associated Nxith power-assertive styles of
Also, studies investigating the
socialization features associatedwith low socioeconomic status highlight harsh or coercive

discipline,
lack of maternal warmth, lack of social support and aggressive adult models
(Dodge, Pettit & Bates, 1994).
With regard to gender, there were no associationsbetween gender and the children's
four
language
derived
from
ive
MacArthur
Narrati
Cod'
the
the
any
of
scales
expressive
ino
LSystem indicating that the boys and girls were not significantly different to each other in
terms of their language abilities.
There were also no significant associationsbetween genderand the four factor scales,
however, the post hoc analysis revealed several significant findings when the individual
As
battery
to
expected,
was
the
gender.
respect
with
examined
were
stem
story
of
variables
described
boys
that
more instances of physical aggression in
there was a strong indication
Boys
than
also showed more negative atypical responses. physical
their stories
girls.
father.
This
disciplining
II
punishment and representations of a
that there are
differences in the thematic content of boys' and girls' narratives.
is
little
it
there
is
different
the
to
very
The picture
performance variables.
when comes
have
judged
Girls
to
boy
between
given more overall adaptive
difference
the
were
and girls.
to
the
dimension
the
This
boys.
to
child
shows,
which
extent
assess
attempts
than
responses
has
the
the
within
storv
the
embedded
conflict
understood
she/he
words,
thrOLI(-Yh
action or
by
the
the
the
designed
the
to
child
and
provided
resolution
of
It
quality
assess
stern. is also
degree to which the child can addressthe conflict. Insofar as that is the intention, there does
in
definition.
be
bias
However,
looking
to
the
not seem
a gender
inherent
when
more closely
definition
is
bias
dealing
that
the
there
at
it possible
is a
in
with aggressionin the stories and
is
has
been
be
to
aggression
which
shown
it
strongly related to gender. For example,
low
the
at
end of this scale would show evidence of extensive use of defensive
responses
little
denial,
the
task,
or no engagement with
extreme avoidance,
maneuvers such as
use of
aggressiveand punitive responseswhich are senselessand provide no resolution to the story
bizarre
high
highly
Responses
levels
or
atypical
response.
showing
of adaptiveness
and
include addressing the conflict, cooperation among characters,richly embellished and sense
high
level
does
The
of adaptiveness
of creative playfulness with minimal sins of punishment.
developmentally
in
into
that
the
take
are
adaptive uses of aggression play
account
not
in
For
for
this age group.
example, a parent may use aggression a story in order
appropriate
is
is
This
danger.
from
that
the
to
parent actually
suggest
not
to protect the child
potential
for
is
the
but
the
and
child
might
positive
the
that
aggression
of
is
meaning
aggressive
kept
being
feelings
safe.
of
represent
is
The one surprising result concerning the performance variables that girls seemed
boys
Is
that
the
are enacting
to show more themes of anger in their narratives. it possible
the
that
the
girls
are
putting
and
aggression
anger through the representations of physical
?
It
intonations
of
course
is,
and
words
as
angry
such
aggression into representations
large
that
dealing
spurious
results
of
variable
group
a
relatively
with
that
when
possible,
be
type
error.
one
a
this
that
could
and
occur
\a(. 0
In conclusion, a central result concernedthe lack of marked associationsbetween the

four factor scales and the demographic data. However, one
main association from the
demographic data to the Discipline/Punishment factor is noted. The effect for father's
social
class on the Discipline/Punishment scale will have to be controlled for in subsequent
analyses.
A second important finding relates to the fact that previous studies have controlled
fIor both childrens age and language ability in their analysis of the MacArthur story stems.
In this instance there were no associations between either of these measuresand any of the
MacArthur scales and as such, subsequent analyses will not control for these.
Post hoc analysis of individual variables indicated differences in the boys and girls
for
in
be
future
therefore,
controlled
where necessary
will
analyses.
responsesand,
The next chapter investigates the concurrent associations between the MacArthur
story stems and a measure of child adaptation.
1 '17
CHAPTER FIVE
ASSOCIATIONS
BETWEEN THE CHILD BEHAVIOUR
THE MACARTHUR
CHECKLIST
STORY STEM BATTERY
AND
128
5.1 INTRODUCTION
So far this thesis has addressed important questions regarding psychometric

issues of inter-rater reliability and the reduction of data in the creating of four factors
trom the many scales of the MacArthur Narrative Coding System (see Chapter Three).
That chapter was confined to a within the MacArthur Story Stem Battery and MacArthur
Narrative Coding Scheme analysis in order to provide descriptive data of the measures.
The previous chapter, Chapter Four, began the exploration of the two measures by
demographic
data
the
the
to
extent
which
particular
was associated to the
examining
follows
in
What
the current, and subsequent,chapters is the application
task.
narrative
its
task
the
and associated coding scheme to several studies to examine the
narrative
of
external validity and to explore possible correlations with other measures, particularly
between
however,
First,
the
the
this
associations
those of attachment.
chapter concerns
Child
the
task
to
the
and
well-established
narrative story stem
Behaviour Checklist (CBCL; Achenbach, 1983), as completed by mothers and fathers.
Relatively recently researchersin the field of psychology have begun to examine

in
behaviour
problems children younger than six years of age and to
the significance of
difficulties
long-tenn
(Campbell,
consider these problems as potential indicators of more
1995). The nature and extent of developmental changes occurring during toddlerhood
for
to
the
children
set out on either a positive
potential
underscore
years
the
preschool
and
develop
be
&N-clopmental
to
that
or
course
adjustment
problems
may
or adapti\c
A
long-standing.
growino number of prospective evidence
transient or
that
in
identified
the preschool years often persist and that children
behaviour problems
129
identified
as showing relatively serious disruptive behaviour problems in earlN-
adolescence often have a history of problems that began in the preschool years
(Campbell, 1990).
Obtaining
self-reported information
from young children has also been
problematic and understanding the child's experience is important for comprehensive

diagnosis of problems. The story stem battery may help clinicians to learn more about the
thoughts and feelings of young children which might permit identification of both
internalising
behaviourdifficulties (Warren,Oppenheim&
externalisingand, especially,
Emde, 1996). In a study of four- and five-year-old children from a non-clinical sample,
between
children's representations of mothers in their play narratives and
associations
Child
Checklist,
Behaviour
the
adaptation,
including
were
measuresof socioemotional
found (Oppenheim, Emde & Warren, 1997). Children, both at age four and five years,
in
by
their
their
their
play narratives were rated
mothers positively
who represented
had
behaviour
fewer
having
problems and children who
more
externalising
mothers as
five
having
four
and
were rated as
more
negative representations of mother at age
four-yearFurthennore,
behaviour
representationsof mothers of
problems.
externalising
four
behaviour
five
behaviour
problems at
years
years and
problems at
olds predicted
had
five
Also,
children who
years.
more
predicted representations of mothers at
disciplinary representations of mothers were rated as having fewer externalising
behavIOUrproblems age four and five.
identify
difficult
to
internalising problems are especially
in young children and,
bulk
focuses
been
the
have
of evidence
on young children with
rarely studied,
therefore,
130
disorder
deficit
disorder
(Campbell.
1990).
and
attention
symptoms such as oppositional
In the study mentioned above, several associations were found between the Internalising
behaviour scale and the story stems. Children who had more disciplinary representations
fewer
behaviour
four
had
five
intemalising
problems at age
and children ,N-hohad
at age
four
had
fI
ewer
more positive representationsat age
behaviour problems at
five.
be
These
to
associations
were
shown
unrelated to mother's psychological
age
distress biasing their CBCL reports and were not mediated by the children's vocabulary.
Associations between children's representations of mothers and their CBCL

however,
the
to
the
pattern of
gender,
children's
examined
with
regard
also
were
scores
has
been
for
both
boys
As
found
be
there
the
to
shown
and girls.
same
associations was
individual
in
difference
to
be
the
variables
to
present study with respect some
a gender
in
Four),
the
Chapter
(see
present chapter.
consideration
merits
gender
such as aggression
Therefore, gender will be controlled for with respect to these variables and the narrative
for
boys
for
be
girls.
and
separately
considered
will
responses
CBCL
Is
between
found
been
the
have
and children
Associations
not only
the
scheme,
coding
narrative
of
the
portion
representation
parental
responses on
between
found
themes
been
have
in
preschool
and
emotions
particular
associations
&
Emde,
Oppenheim
Checklist
(Warren,
Behaviour
Child
the
and
children's narratives
in
distress
five-year-old
fourthe narrative
children expressing
and
1996). For example,
both
found
to
task
telling
correlate
positively
with
mother's
and
was
the
story
of
phase
tIather
destructive
CBCL.
Aggressive
the
inos
and
the
of
scale
externalising
oti
rat ,II
found
five-year-old
fourto
also
children
Nvere
correlate
and
of
the
narratives
thenics in
1311
significantly on both the externalising and internalising scales of the CBCL as rated by
father
mother,
and teachers. Aggressive and destructive themes included aggression,
injury
and atypical negative response. Another study using the story stems has
personal
compared samples of maltreated and non-maltreated children and found that the
in
themes
their narratives
maltreated childrens' narratives contained more aggressi,,,,
e
(Buchsbaum, Toth, Clyman, Cicchetti & Emde, 1991).
The majority of the few published studies exploring the associationsof childrei-i's
by
narratives
and
measures
of
socioemotional
adaptation
utillse ratings
play
mother
(Bretherton, Ridgeway & Cassidy, 1990-, Oppenheim, Emde & Warren, 1997:
Oppenheim, Nir, Warren & Emde, 1997). There has been only one to report on the
Oppenheim
&
Emde,
(Warren,
1996).
The
father's
the
child
rating of
associations with
father's
for
that
and
ratings,
mother's
is, the
appearing
profile
authors report a similar
This
in
the
their
child.
study will also
to
of
ratings
parents tended agree with one another
in
father's
between
to
relation
and
ratings
the
mother'
similarities
of
question
examine
the child narratives.
Overall, this chapter aims to replicate the suggestion that emotions, themes and
be
in
the
assessment
useful
in
could
narratives
play
children's
representations
parental
MacArthur
task
the
the
narrative
provide
could
question,
asks
and
problems
of childhood
be
for
to
thought
to
too
young
reliabl
ordinarily
children
measure
self-report
a useful
behaviour
?
their
problems
discuss the presence and extent of
be
factor
ling-iNegative
Control
the
will
positively associated
it is predicted that
132
with either/both the externalising and internalising scales as reported by mother and
father. It is further predicted that the factors of Discipline/Punishment, Quality/Open
Responseand Positive Maternal Representations will be negatively related to both the
internalising and externalising scales.
5.2 METHOD
5.2.1 Design
As has been previously described, the third phase of the London Parent Child
Project consisted of two visits by the families. The first visit took place as close to the
fifth
birthday
later.
during
It
the
the first
as possible and
second six months
child's
was
Stems
data
for
Child
MacArthur
Story
Behaviour
Three
Phase
the
that
the
the
and
of
visit
Checklist, to be discussed in this chapter, was collected. While the child subjects were
language
Story
Bus
Story
Stem
Battery
MacArthur
the
the
assessment
and
completing
(described in Chapter Three and Four) in the play hut at The Anna Freud Centre the
for
in
building
to
them
to
the
taken
complete various
away
order
yards
main
was
parents
Checklist.
Behaviour
Child
including
the
assessments
5.2.2 Subjects
The subjects are those of the London Parent-Child Project that have been
described in the preceding chapters.For the third phaseof the study the 100 families were
fifth
birthday
in
before
follow-up
the
to
the
and
child's
asked
participate
contacted shortly
boys)
61
(43
49
(range
59the
The
children
girls
was
of
and
months
age
mean
research.
I l 3
65 months). Eighty-nine children completed the MacArthur Story Stem Battery and each
father
mother and
completed a Child Behaviour Checklist. Overall, there were lo%\
baselines for problem behaviours whether rated by mother or father, wliich is to be
low
the
given
expected
risk nature of the sample, and consequently very feN-childreti
borderline
the
were rated in
clinical or
regions.
5.2.3 Procedure
Child Behaviour Checklist (CBCL)
As the children were involved in. completing the MacArthur Story Stem Battery
hut
house
Story
Bus
their
the
the
the
parents were waiting in
main
play
where
in
and
form
Child
Behaviour
Checklist
fill
the
the
to
of
parental report
out
each was asked
(Achenbach & Edelbrock, 1983) (Appendix Q. Each parent filled out a separateform
between
there
partners.
was no conferring
and
The Child Behaviour Checklist forms part of a multi-axial empirically based

functioning
(Axis
IParental
Reports,
behavioural/emotional
assessmentof children's
Axis 11 - Teacher Reports, Axis V-
Direct Assessments of the Child), with high
Revised
Behaviour
Problem
Conners
the
the
and
scales
concurrent validity with
discriminant
distinguishing
between
have
been
had
to
It
Checklist.
validity.
shown
also
important
&
Edelbrock,
feature
1983).
(Achenbach
An
children
non-referred
and
referred
is
Checklist
the
Behaviour
Child
narrow-band
syndrome
scales
it
produces
as
part
the
of
for
behaviour
(4-5
6-11
sex
age
profile
each
at
specific
points
yrs.
child
standardized
a
of
134
yrs, and 12-16 yrs) derived from performing separatefactor analysesfor children of each
sex and each age group (N=2300) who had been referred to the mental health service in
the United States.Each of these scaleshas been normalised by comparison with CBCLs
by
1300
completed
parents of
randomly selected 4-16 year olds who had not been
health
to
the
referred
mental
services in the previous year. The narrmN-band scores fall
description
the
under
of broad-band problem scales, internalizing, externalizing and
mixed.
5.3 RESULTS
The results of the data analyses of the Child Behaviour Checklist and the
MacArthur Story Stems are reported in three sections. The first section addressesthe
level of agreement between the mother and father ratings on the two scales of the Child
Behaviour Checklist. The second part of the results consider the data obtained from
Checklist.
The
Child
Behaviour
the
the
secondsection
subject child on
mother's rating of
for
father's
data
the
the
the
the
of
child
rating
on
analyses
of
results
with
is concerned
Child Behaviour Checklist. In both sections gender is controlled for where appropriate.
5.3.1
Agreement between maternal and paternal reports
father's
Child
both
the
In the one studN,,,
ratings of
mother's and
which reported
Behaviour Checklist as associated with children's narratives, the authors state that the
(Warren,
Oppenheim
the
their
of
child
other
in
rating
to
each
with
tended
agree
parents
r-I
if
discover
In
the parents in the current sample tended to agree
to
1996).
order
& Fnide
1-s-
or disagree more with each other in their view of the child a correlational anal.ysis was
done comparing their ratings on the two scales. Table 5.1 presents the results of the
Pearson correlations for mother's and father's ratings on the externalising and
internalising scales of the Child Behaviour Checklist.
Table 5.1 : Correlations of mother's and father's rating on the CBCL scales (n =92).
father's externalising
mother's
mother's
externalising
intemalising
446 (. 000)*
.
father's intemalising
father's total
Note: p :! 00 1
.
mother's total
197
(.06)
506 (.000)*
.
(p values are based on two-tailed tests and are in parentheses)
As can be seenin the abovetable, motherand fatherhave strongeragreementon

While
internalising
for
the correlation on the
the
than
scale.
the externalising scale
highly
high
the
the
correlation
on
significant,
and
moderately
externallsing scale is
father
is
low
tend
to
the
that
and
rate
child
mother
and
indicates
scale
internalising
differentl-y from one another for internalising behaviour problems. As a result, it was
father's
for
be
These
discussed
do
decidedto
ratings.
will
separateanalyses mother's and
beginning
following
two
mother's
with
ratings.
sections
the
in
1336
5.3.2 Mother's ratings on the CBCL
Table 5.2 below gives the Pearson's correlations for mother's ratings on the
internalising behaviour scale, the externalising behavior scale and a combined or total
CBCL.
the
score of
Table 5.2: Correlations and levels of significance of the four story stem factors with
CBCL.
the
on
mother's rating
Quality/open
Discipline/
Controlling/
response
punishment
negative
Positive maternal
representation
07)
070
53)
194
(.
(.
177 (. 10)
058 (. 58)
-.
.
.
.
036 (. 74)
189 (. 07)
070 (. 52)
014 (. 89)
Externalizing
.
.
.
.
193 (. 11)
149 (.22)
103 (.40)
044 (. 72)
Total
.
.
.
.
Note : r-values are Pearson correlations; levels of significance are in parentheses;
05
10
between
trend
treated
a
as
are
values
-.
.
Internalizing
level
the
5.2
from
Table
As can be seen
above, there are no significant results at
for
behaviour
intemalising
for
There
two
one
and
three
05
results,
are
above.
or
of p =.
for
is
is,
That
trend
the
indicate
there
trend.
behaviour,
a
that
a
externalising
factor
(r = 177, p=.
.
Representation factor (r = . 194, p=

behaviour
problems
Controlling/Negative
Z--
as
10) and the Positive Maternal
07) to correlate positively with internalising

.
for
There
by
trend
is also a
mother.
rated
factor (r = 189, p=
.
behaviour
problems.
externalising
the
07) to be positively associated with

.
13
5.3.2a Controlling
for father's social class
Although mother's rating on the interrialising scale correlates just at the level
of
for
trend
the Discipline/Punishment factor it was decided that a cautious approach
a
be
in
taken
would
relation to father's social class as it proved to have a sionificant main
for
the Discipline/Punishment factor in Chapter Four. Therefore, father's social
effect
class was controlled for
in a partial correlation analysis. The correlation remamed
level
the
significant at
of a trend (r = .216, p =. 08).
5.3.3 Post hoe analysis
As the trends found above might prove to yield informative results if examined
investigate
decided
to
those areas where there are trends in more
more closely, it was
detail. The individual variables comprising each factor were looked at in relation to either
the internalising or externalising rating, as indicated above in Table 5.2, to see if there
was any one particular variable or variables which prove significant.
The Dicipline/Punishment factor is comprised of the variables excludes other,

disciplining
father.
and
mother and
physical punishment, shameother, verbal punishment
The table below (Table 5.33)presents the Pearson correlations of the mother's rating of
these
variables.
of
each
the child \\ ith
138
Table 5.3: Correlations of mother's rating of the child on the internal isi
scale
il
(T
with
fqc-. tnr
variables comi)risinLyDiscinline/Punishment
Internalising
1
excludes
other
physical
punishment
shame
other
verbal
punishment
disciplining
father
mother
151
.
(. 16)
097
-.
(. 37)
32 8
.
(. 002)**
165
.
(. 12)
024
-.
(. 82)
1433
.
(. 18)
Note : r-vaiues are Fearson correlations; levels of'significance

** p :! 01
.
disciplining
are in parentheses
An interesting result emerges from this 'unpacking' of the factor scale. The
variable 'shame other' appears to be the one variable that Is sIgnificantIly related to the
internalising scale as rated by mother.
The second factor that indicated a trend on the intemalising scale is the Positive
Maternal Representation factor. The two variables of affection and positive maternal
factor
Pearson
the
this
and
correlations of these with mother's
representationconstitute
CBCL
below.
internalising
I
the
the
the
scale of
are given
child on
rating of
Table 5.4: Correlations of mother's ratings of the child on the internalising scale with
factor.
Representation
Maternal
Positive
variables comDrisina
affection
Internalising
109 (. 31)
positive maternal representation

240 (. 02)*
.
11
Note : r-values are Pearson correlations, levels of significance are in parentheses

*p !. 05
9
1
From the above result, it appears that positive maternal representation the
is
variable that is significantly related to the internalising scale.
Only the Controlling/Negative factor related to mother's rating of the

child on
the externalising scale. The Controlling/Negative factor is comprised of negative atypical
response,p ysica aggression,investment in performance, control and a lack of anxiety.
Table 5.5 below presentsthe Pearsoncorrelations of thesevariables with mother's ratings
on the externalising scale.
Table 5.5: Correlations of mother's rating of the child on the externalising scale with
I
ling
factor.
Negative/C
ontro
xte nalising
negative atypical
response
physical
aggression
investment in
024
.
(. 82)
259
.
(. 01)**
076
.
(. 47)
control
lack of
anxiety
149
.
(. 16)
173
.
(. 10)
performance
Note : r-values are Pearson correlations; levels of significance are in parentheses

** p : 01
.
Again,
it appears that there is one variable, physical aggression. v,'hich is
for
lack
There
trend
the
is a
of
externalising scale.
associated with mother's rating on
be
the
to
externalising scale.
with
associated
positively
anxiety also
140
5.3.3a Mother's ratings and controlling for gender
In Chapter Four it was indicated that certain variables are significantly related to
Of
these, physical aggression is the variable relevant to the analyses beino
gender.
in
this section. In order to examine the issue of gender for this variable in
conducted
CBCL
to
relation
scores,, correlations were conducted for mother's ratings on the
externalisingscalewith physicalaggressionfirst for only the boysin the sampleand then

(See
Table
below).
5.6
with girls.
Table 5.6: Correlations of physical aggression and the externalising scale as rated by
for
boys.
2irls and
mother
physical aggression
Externalising
girls
(n= 40)
boys
(n= 46)
097
.
(. 55)
264
.
(. 07)
Note : r-values are Pearson correlations; two-tailed levels of significance are in

parentheses
It can be seen that physical aggression is positively associated with mother's

is
level
(p
07)
boys
for
trend
the
the
of a
=.
and
at
only
rating on the externalising scale
for
girls.
not associated at all with mother's rating
141
5.3.4 Father's ratings on CBCL
The same analyses as was conducted above for mother's ratings was conducted
for father's ratings and is reported below. Table 5.7 gives the Pearsoncorrelations for
father's rating of the child on the internalising behaviour scale, the extemalisim,
behaviour scale and the combined score of the Child Behaviour Checklist.
Table 5.7: Correlations and levels of significance of father's rating of the child on the
CBCL for the four storv stem factors.
Quality/open
Discipline/
Controlling/
response
punishment
negative
Internalizing
102
34)
(.
-.
Externalizing
[i
28)
114
(.
-.
I
038 (. 72)
.
063
(.
56)
-.
060
(.
57)
-.
206 (.05)*
.
=
023
84)
(.
-.
Positive maternal
representation
051 (. 63)
.
001
98)
(.
-.
(.
37)
7
(.
82)
111
(.
25)
140
-.
-.
*p:!
.05
There is one significant correlation between the Controlling/Negative factor and

have
Children
and
more negati%,
e
the externalising scale.
who are more controlling
behaviour
have
father
by
in
problems.
more externalising
themes their narrativesarerated
-1
14
5.3.5 Post hoe analysis
As with the mother data, it was decided to investigate the area

where there is a
significant
result
in
more
detail.
The individual
variables compr IIs n-o the
Control ling/Negative factor were looked at in relation to the

externalising rating, as
indicated above in Table 5.7, to see if there was any one particular variable or variables
which prove significant.
Table 5.8 below presents the correlations with each of the
variables that comprise the Negative/Controlling factor, that is, negative atypical
response, physical aggression, investment in performance, control and lack of anxiety.
Table 5.8: Correlations of father's rating of the child on the externalising scale with
factor.
Ne2ative/Controllin
variables comprisin
xternalising
negative atypical
response
physical
aggression
investment in
037
-.
(. 73)
200
.
(. 06)
084
.
(. 43)
control
lack of
anxiety
186
.
(. 08)
212
.
(. 04)*
performance

*p:! ,
.05
Rather than one highly significant variable emerging, Table 5.8 shows that there
found
three
to
the
variables
contributing
of
significant result
above
is a combination
(Table 5.7). A lack of anxiety is positively associated with externalising behaviour
father.
by
Plivsical
ag
and control are also related to
problems as reported
-gression
level
behavIOLir
the
trends.
at
problems
of
externalising-,
14 l
5.3.5a Father's ratings and controlling for gender

Physical aggression correlated with father's rating on the externalising scale is
is
being
there
trend
towards
although
a
non-significant,
it
significant. As such, it was
decided to investigate further by doing a separate correlational analysis for physical
boys.
Table 5.9 below presents these figures.
then
aggression with girls and
Table 5.9: Correlations of physical aggression and the externalising scale as rated by
father for girls and boys.
physical aggression
Externalising
girls
(n= 40)
boys
(n= 46)
013
.
(. 93)
175
(.24)
Note : r-values are Pearson correlations; two-tailed levels of significance are in

parentheses
for
findings
figures
5.8),
(Table
from
the
be
As can seen
the above
mother,
unlike
boys
is
to
is
or girls.
there no significant result when physical aggression correlated
5.4 DISCUSSION
between
the children's
the
This chapter was concerned with
associations
Checklist.
It
Child
Behaviour
the
task
shed
and
well-establi
the
to
narrative
responses
from
findinos
that
themes
that
emotions.
and
suggested
studies
several
to
aimed replicate
be
the
assessment
stories
could
useful
of
play
in
children's
in
parental representations
childhood problems.
144
Like the one studied reported in the introduction, there was

general agreement
between the parents' ratings on the scalesCBCL. The agreementbetween
parents' ratings
for the externalising and total scales was high. However, there
less
vas
agreement
between parents on the internalising behaviour scale.
The findings from the association of mother's ratings to the four

story stem
factors are somewhat disappointing. There were no significant associations
and it was
expected that a clearer relationship would emerge particularly between the Positive
Maternal Representation and Control Ii ng/Negative factors. As mentioned,
internalising
problems are particularly difficult to identify in young children. However, the decision
to look more closely at the component variables of theses factors proved interesting. In
the case of mother's ratings, it was discovered that the one component variable which
highly
interrialising
to
the
was
significantly related
scalewas 'shame other'. Shameother
is
if
incident
involving
the
content variables and scored present or absent an
is one of
is
For
example, one charactermay say to another.
shaming another character mentioned.
'Shame on you, you are bad !' or 'You are a naughty girl ! '. This relationship might be
displace
by
that
theory
suggests
children
often
which
psychoanalytic
or project
explained
feelings of
in
their play.
unworthiness or shame onto other people or characters
Unfortunately, the variable describing 'shaming self was not reliably rated and so is not
frequency
Nevertheless,
for
the
of times the themes are used
comparison.
available
100
in
'shame
that
this;
times
might
support
other'
used
over
evidence
is
provides some
the storics and 'shame self- only eight.
As expected.
there was a relationship between mother"s rating on the
145
internalising scale with the positive maternal representation variable, however, the
direction of the relationship was not expected. Those children who had more positive
had
mother's NA-horeported more internalising child
maternal representations also
behaviour problems. This is contrary to the above mentioned study and perhaps counterintuitive. However, two possible explanations come to mind. It is possible that children
who are experiencing internalised problems receive more praise and positive
from
their mother and so come to represent their mothers as positive
encouragement
intemallsed
having
be
It
that
children who are rated as
may also
reflecting reality.
behaviour problems idealize their mothers and defend against any painful perception of
by
benevolent
is
the
representing a caring
child wishes
as
not as
a mother who
is
It
judgements
data
have
do
this.
We
to
the
about
make
not
mother.
affectionate
is
It
home
in
possible
always
the
evidence.
such
provide
that
might
observations
possible
that the result is a spurious, type one error.
between
found
physical aggression,
Not so surprisingly, there was an association
factor,
I
ing/N
Control
rating
mother's
and
the
e
gative
one of the component variables of
behaviour
having
Those
externalising
as
rated
children
on the externalising scale.
This
in
true
the
after
even
was
instances
stories.
had
aggressin
of physical
more
problems
for
gender.
controlling
father's
the
between
the
on
child
of
ratings
There was one significant relationship
Those
factor
ling/Negative
Control
who
children
as
predicted.
the
and
scale
externalisim,
have
to
lino
Control
/Negative
externalising
reported
were
scale
l"(1her
the
on
scored
do,,
broken
factor
Vhen
component
variables
Its
Into
N-n
this
was
behaviour problems.
146
it was shown that there was no one variable strongly associated to father's ratings but
rather three variables seemedto be responsible for the overall significant result obtained
for the Controlling/Negative factor. These were physical aggression, control and a lack
of anxiety.
Physical aggression in the stories is distinct from both verbal aggression and
physical punishment. For example, it is scored when the subject has one doll push or hit
doll
another or one
pushes another off a bike or chair. Control is one of the performance
indicated
by
the subject attempting to control the examiner or the
and
variables
is
For
situation.
example, the child may refuse to relinquish control of a doll or prop. The
definition of control includes the understanding that the subject child must persevere in
their controlling attempts, that is, just one incident of refusing to give up a doll would not
be counted. The lack of anxiety associatedwith behaviour problems on the exterrialising
in
both
is
that all of these
to
physical aggression and control
possibly related
scale
feelings
into
Anxiety
be
action.
would
variablesmight seenas ways of putting negative
face,
hand
lips,
body
defined
(repetitive
be
rubbing
chewing
movements,
coded as
not
helplessness
in
feelings
be
because
their
)
the
of
anxiety
and
enacting
children
would
etc.
in
both
behaviour
the
the
and
with
examiner.
play
aggressive
controlling and
In summary,therewere few major differencefound betweenthe associationsof

The
father's
the
one significant result of
child narratives.
ratings and
mother's and
tathei's rating on the externalising scale being related to the Controlli
factor
for
by
the
trend
result
mother's
althoug
not
statistically
the
same
in
rating,
is echoed
,h
s"In
1. Ificant.
di\
the
I n -idual variables are compared
similar piicturc emerges when
th
147
In
is
father's
the
case
of
mother's
mother's and
ratings, it aggression which is
ratings.
the telling variable and in father's caseit appearsto be a combination of a lack of anxiet,,
This
seems to be confinning the results described by
physical aggression and control.
(Warren et al. 1996),,however, caution must be observed in interpreting the above result
as none of the results are particularly strong. This could be due to the nature of the nonclinical sample.
The following
chapters will
investigate the concurrant and longitudinal
between
the narrative task and measuresof attachment organisation.
associations
148
CHAPTER SIX
ASSOCIATIONS
BETWEEN THE MACARTHUR

AND CHILD-PARENT
STORY STEM BATTERY
ATTACHMENT
149
6.1 INTRODUCTION
The previous chapter explored links between the story stems

and the child's
behaviour
adaptive
and attempted to replicate the few previously published studies
relating to representations of the child's internal world to socio-emotional functioning.
This chapter moves the investigation of the young child's internal

world into
the area of attachment theory and investigates the possibility of concurrent
between
the children's responsesto the MacArthur Story Stem Battery
associations
and their attachment security as measured by the Modified Strange Situation.
As has been suggestedearlier in this thesis, attachment theory is of crucial

discussing
link
between
the
the parent-child relationship and the
importance when
developing
in
the world outside of that relationship.
child's
understanding and skills
Early attachment theory proposes (see, for instance, Bowlby, 1973) that an
later
become
'transformed'
into
early
attachment
relationships
individual's
inner
have
implications
for
these
that
representations
predictable
representations, and
in
later
life.
experiences, particularly emotional ones,
In a study conducted with three- and four-year-old children and their parents,
discovered
between
the children's responses on the
were
concurrent associations
MacArthur Story Stem Battery and measuresof family relationships and adaptation
&
Winfrey,
1993').
In
fifty-one.
Emde
a
sample
of
(Oppenheim,
primarily white
found
families,
that
the
they
children's narratives about affect and
middle-class
1510
in
conflict themes were associated predictable and theoretically meaningful ways with
by
the socio-emotional adaptation of children and parents as measured the CBCL (see
family
In
the
this case, the
quality of
relationships.
previous chapter) and with
in
family
differences
functioning
in
the
quality of
children's narratives
expression of
depended on gender. Narrative coherence served as the Indicator of familY
functioning for boys and aggressive themes for girls. Aggressive themes included
aggression, verbal conflict, escalation of conflict and atypical negative responses.
This informs us about a link between representations in the stories and general family
functioning as measured by a parent-child relationship questionnaire. A limitation of
domains
did
it
is
themes
that
all
not include
the study
of conflict and
concentrated on
System.
Coding
MacArthur
of the
Contemporary attachment work seeks empirical investigation of the link

later
later
quality of
between early quality of attachment,
quality of attachment, and
The
1991).
Grossmann,
to
&
Grossmann
adds
for
chapter
present
life (see
example,
between
this investigative effort specifically with respect to concurrent associations
Strange
Modified
by
(as
the
in
individual differences
assessed
attachment security
by
MacArthur
(as
the
in
differences
task
assessed
Situation) and individual
a narrative
in
discussed
literature
this
the
This
to
earlier
Battery).
Stem
Story
study contributes
Ston,
MacArthur
the
the
lending
by
to
validity of
support
evidence
thesis primarily
look
by
literature
the
taking
at
a
unique
to
Stem Battery, and adds attachment
the
and
preschool
years
preschool
between
in
the
relationship
attachment
connection
literature
for
territory
as
attachment
ILinctioning. This specitic connection is new
12
18
beyond
behaviour
(the
to
months
for
of
attachment
patterns
assessing
methods
15
'Ainsworth system,' discussed in the next chapter) are recent developments. A systern
for two and a half- to four and a half-year-olds was developed in 1989/1992 by
Cassidy and Marvin, and a system for six-year-olds was developed in 1988 by Main
Cassidy.
The
and
assessmentof patterns of attachment behaviour in this study is
Cassidy
the
made using
and Marvin system.
There has been only one report in the literature concerning concurrent
Cassidy
the
of
attachment
assessment
using
and Marvin (1989/92) Modified Strange
Situation and a story stem task (Bretherton, Ridgeway & Cassidy, 1990). The study
five
three-year-old
conducted with
children used
story stems from the Battery, three of
However,
the
to
the
the stories were not
are
same
administered
present sample.
which
developed
MacArthur
Coding
Scheme,
the
rather they were
recently
coded using
disorganised
demonstrating
relationships.
secure, avoidant, resistant or
categorised as
These were then compared to the attachment classifications of the Modified Strange
Situation. The concordance of secure versus insecure classifications for both
insecurity
however,
the
type
of
was not consistent across
procedures was significant,
found
The
that the story scores were also significantly related to
authors
procedures.
but
18
Situation
the correlations were not as
Strange
the
months,
with mother at
between
the
storey scores and concurrent separation-reunion
association
strong as
security.
itself
limited
to assessingthe stories with an eye to
The study above
Codino System is a more systematic approach and covers
MacArthur
The
attachment.
in
This
to
that
allows
naturally
children's
play.
us
occur
explore
thenies and content
I -)
themes or groups of themes that might seem unrelated to attachment at first glance but
further
inform
links
between the child's concurrently assessed
which might
about
us
his
her
functioning.
This in turn might lead to further
classification
and
attachment
or
understanding of the manner in which attachment is organised in five year olds -a
relatively unexplored area.
6.2 METHOD
6.2.1 Design and Subjects

The participants discussed in this chapter are a part of the London ParentChild Project sample, which has been referred to throughout this text. The data
from
longitudinal
is
in
the
this
this
most recent stagesof
project
chapter
presented
in
life.
five
At
the
two
sixth year of
years,
visits when participants were
collected on
the adapted Strange Situation procedure (Cassidy & Marvin, 1992) was conducted
lack
Through
this
the
sample was
at
visit,
or
relocation
response
of
mother.
with
boys)
89
46
leaving
(43
by
7%,
from
the original sample
girls and
children
reduced
five.
five
At
Situation
Strange
in
and
with mother at age
who were seen the modified
Strange
Situation
"sixth
(referred
the
half
the
to
to six years
modified
year visit")
as
a
The
in
father
'strange'
sample was reduced
playroom.
a new and
\vas conducted with
by a further 20% and 71 children (35 boys and 36 girls) were seen at this time with
fIather.
Of the 89 children who were assessedin the modified Strange Situation
five
57
(B),
19
their
were
classified
as
secure
at
\-cars,
mothers
as
with
procedure
7 as insecure-ambivalent (C). and 6 as di
(A).
IIISCCLire-avoidant
(D). Of the
1
I 'j
71 participants who attended the second visit and were assessedat this time in the
Strange
Situation with father (46 secure, 21 avoidant, and 4 ambivalent),
modified
No children were classified in any of the insecure.

disorganised
controlling or
groups with father.
6.2.2 Procedure
6.2.2a The Modified Strange Situation

The Modified Strange Situation procedure is an variation of the original
system developed for use in infancy (Ainsworth et al., 1978), and the system
developed for use with older children (Main & Cassidy, 1988). The adaptation, most
from
in
'stranger'
the
the
eliminates
procedure, an effort to make it
importantly,
for
first
This
by
preschoolers.
appropriately stressful
version was
used Marvin and
Van Devender (1978) and is recommended by Cassidy and Marvin (1989/1992) for
from
involving
Apart
'stranger',
the
the
eliminating
preschoolers.
assessments
role of
the procedure remains the same as the Ainsworth's infant Strange Situation. A
follows:
Strange
Situation
Modified
the
summary of episodes of
Episode 1 (3)minutes):
Episode
-'
(3)minutes)*
remains alone.
Warm up, parent and child are in the playroom together.
Sei)(wation 1, parent leaves the room and the child
154
Episode 3 (3 minutes):
Reunion 1, parent returns.
Episode 4 (6 minutes)*:
Separation 2, parent leaves the room and the child
remains.
Episode 5 (5 minutes):
Reunion 2, parent returns.
* These episodes are ended or shortened if child becomes unduly distressed
6.2.2. b Coding Procedure for Modified Strange Situation

The system used to classify and rate organisations of attachment behaviour in
basis
described
five
the
the
of
procedure
above involves
main
preschoolers on
descriptions
found
Full
be
two
can
rating scales.
in the
attachment classifications and
is
&
1989/1992')
Each
by
Marvin,
(Cassidy
classification
raters
coding manual used
briefly described below.
Attachment Classifications
Secure
I.
interactions
intimate
immediately
with
calm upon reunion,
very secure: nearly
full gazeand positive affect. The interactionshave intent and indicate that the
is
special;
relationship
initial
reserve upon reunion
secure-reserved: compared with very secure,
less
(subset
interest,
less
I
iii). or
visual and/or
(subsets and ii),
openness and ease
(iv);
thouoh
responsive
verbally
even
proximity,
physical
3.
behaviour.
but
secure
elements
of
either
generally,
resistant
secure-ambivalent:
1
immature
behaviour
also present;
or
behavlour,
but
for
generally
secure
either a mild struggle
secure-controlling:
for
behaviour;
the
traces
of
reunion
episode,
or
of
controlling
much
control
5.
behaviour,
but
does
fall
generally
secure
secure-other:
either
not
into any of
the above groups, or shows a mixture of groups.
Insecure-Avoidant
I.
ignoring: extreme physical and conversational avoidance, minimal responses;

less
extreme avoidance.
neutral:
Insecure-Ambivalent
I.
directed
to
to
or
parental suggestions;
parent
whiny
resistance
resistant: angry,
2.
immature: immature, coy behaviour, often with ambivalence to physical
proximity/contact.
Insecure-ControllingIDisorganised
I.
caregiving:
'cheerful' behaviour, perhaps in an overly 'bright' greeting or show
interactions;
by
followed
but
congruous
and
affect
positive
toys,
up
not
of
2.
in
do
behaviour,
to
hostile
telling
a
what
parent
perhaps
punitive,
punitive:
derogatory, annoyed or angry manner;

but
forms,
the
both
of
the
control
clear
above
or neither of
general:
interactions;
4.
incomplete
disordering
or undirected
disorganised:
of expected sequences;
dazed
behaviour.
and
apprehension,
confusion
stereotyped
including
moveiiients,
cxpressions,
dcpressed
affect.
and
Insecure-other
156
I.
insecure behaviour with either does not fall into any of the above insecure
groups, or shows a mixture of groups.
6.3 RESULTS
The results are divided into two sections: the first section describes the data
in
relation to the child's attachment to mother at age five using the Modified Strange
Situation. The second section presents the results for the child-father attachment data.
6.3.1 Child-mother
data
Independent sample t-tests were calculated to determine if there were

four
between
factors
derived
from
the
the children's responses
story stem
associations
to the narrative task (see Chapter Three) and their concurrent attachment classification
Strange
Situation.
Table
6.1
below
by
Modified
the
presents
with mother as assessed
for
four
levels
deviations
the
the means,standard
story stem
and
of significance
factors groupedby insecureversussecureattachmentclassification with mothers.
Levene's test for the homogeneityof variancewas observedand the values for the
equal and unequal groups used accordingly.
157
Table 6.1 : Means and SDS of the four story stem factors grouped by child insecuritv
five
vears of aae.
vs. securitv with mother at
Means (SD)
Factor
QUality/Open Response
Controlling/Negative
Positive Maternal
Representation
df
T-value
Significance*
insecure
(n =32 )
secure
(n=54 )
1.7211 (. 29)
1.8001 (.22)
84
1.40
1793(. 08)
1558(. 10)
81
1.07
.
28
.
84
"2
75
.
81
07
.
94
.
.
5348(. 18)
.
1261 (11)
.
5490(. 22)
.
.
1242(. 12)
I
Note: * two-tailed levels of significance
16
As can be clearly seen from Table 6.1 above, there are no significant
five
factors
four
between
to
the
and security of attachment mother at age
associations
it
deemed
lack
to
As
the
conduct
meaningless
was
associations
of
of
a result
years.
four
investigations
hoc
three
further
the
and
way classifications of
using
post
any
attacliment.
Given the lack of positive findings with the child-mother data, the question
data
?
factors
the
be
the
child-father
with
there
of
associations
will
remains,
6.3.2 Child-father
data
if
determine
there were
to
Independent sample t-test were also calculated
from
derived
factors
four
the
between
children's
responses
the
stem
story
associations
Three)
Chapter
their
(see
concurrent attachment classification
and
task
to the narrative
6.2
below
Situation.
Table
Strange
Modified
by
the
presents
father
assessed
as
with
four
for
levels
deviations
the
story
stem
of
significance
and
the means,standards
father
insecure
by
Ith
as
versus secure attachment classification X',,
factors grouped
15
by
Situation
five
Strange
Modified
the
measured
age
and a half years. Levene's test
for the homogeneity of variance was observed and the values for the equal and
unequal groups used accordingly.
Table 6.2 : Means and SDS of the four story stem factors grouped by child insecuritv
father
five
at
vs. security with
and a half vears of aize.
Factor
Means (SD)
QUality/Open Response
Control ling/Negative
Positive Maternal
Representation
Note:
insecure
(n= 25)
secure
(n= 45)
1.7040 (. 35)
1.8073 (. 16)
1746(. 10)
1596(. 08)
.
5381 (. 15)
.
1206 (. 1
.
I
* two-tailed levels of significance
df
T-value
29.88
1.37
66
.
5328 (.21)
.
1384(. 12)
.
18
.
53
.
91
.
58
.
63
.
11
.
55
.
68
66
I
Significance*
Again, there is no significant result to report. Therefore, there is no point in

these
further
are
addressed
and
are
problems
several
analyses until
performing
discussed below.
6.4 DISCUSSION
This chapter aimed to investigate the possible associations between the

MacArthur Story Stem Battery and a concurrant measure of child-parent attachment,
four
factors
lack
The
Situation.
the
to
Strange
Modified
of
results
of
positive
the
is
data
disappointing.
either the child-mother or child-father
is,
being
be
first
that
is
attachment
asked
adequately
must
One of the
questions
159
assessedby the chosen method ? The childhood attachment classification procedures
have
been developed recently, and must endure reasonable
this
presented in
chapter
speculation as to whether or not they are really measuring attachment (Speltz,
Greenberg, & deKlyen, 1990). Given the vast developmental changes
which occur
from infancy to the preschool years, the use of an attachment classification
system
is
infancy
to
the
which very similar
system must be questioned.
The first question is that of the stressfulness of the Strange Situation for
The
defining
this
children of
age.
element, of course, of the Strange Situation in
infancy is that the infant finds the situation stressful, allowing observers to witness
attachment behaviours of infant to parent as the infant seeks to deal with this stress
(secure infants using parent as a "secure base" for comfort in this stressful time).
Simply becausethe Strange Situation is widely supported as a useful measure of
infancy,
be
assumedthat this remains the suitable context for
attachment in
it cannot
(Cicchetti
1990).
If
et al.,
children of
measuring early childhood attachment as well
this age are no longer stressedby this situation, then one can assumethat measuring
for
is
invalid.
Empirical
in
this observational
support exists
attachment this context
least
As
in
to
mentioned in the
up school age.
method of measurement children at
between
found
(1990a)
Bretherton
a significant relationship
et al.
introduction,,
Cassidy
Marvin
the
and
children
using
system and
attachment security of preschool
In
based
addition,
attachment
the narrative assessmentof attachment.
classification
has
'/2
been
4
the
at
years
associated
situations
with
children's
separation-reunion
on
SAT,
task
representational
involving separation pictures
the
narrative
to
a
response
19922).
The
however.
Stevenson-Hinde.
&
question remains,
with respect
(Shouldice
160
to children in the sixth year. Perhaps this situation is one which a child of this age no
longer finds 'strange' or stressful.
The point was made earlier that representational aspectsof life and language
five-year-old
important
be
This
to
a
child.
may
a shortcoming of the
are extremely
present assessmentof attachment, the reliance on one'type'of
measurement onlY. It
is possible that complete reliance on an observational method of measuring

This
the
this
age.
complexity of children at
is
attachment is not in congruence with
to
that
the
to
suggest
multiple
only
assessment,
it
is
separation-reunion
reject
not
for
be
children of this age, providing investigators with more
useful
assessmentsmay
'clues' than they perhaps need in assessingless 'complex' infants (see Ainsworth,
1990; Main & Cassidy, 1988). With respect to attachment classifications at this
by
developed
have
language
the
these
sixth
children
the
which
sophisticated
stage,
describes
how
both
is
inefficient
it
the
to
and
indicate
child
consider
that
not
year may
how
intentions,
feelings,
the
his/her
thoughts,
and
and
recognises emotion, narrates
behaves.
child
between
the
in
connections
In conclusion,
this chapter no meaningful
five
attachment
child
concurrent
and
age
at
children's narrative responses
The
but
posed
question
is
next
are
posed.
questions
meaningful
made,
classifications
more
offer
relationships
the
attachment
investigating
of
early
quality
is \vIII
?.
by
factors
that
the
and
children's
in
influencing
regarding
meaningful information
find
stronger
confirmation
of
one
will
relationship
attachment
this
early
considering
161
the attachment -narrative link.
162
CHAPTERSEVEN
ASSOCIATIONS

AND INFANT-PARENT
STORY STEM BATTERY
ATTACHMENT
1633
7.1 INTRODUCTION
Chapter Five found associations between the story stems and a vellestablished measureof child adaptation, the Child Behaviour Checklist. The pre'ious
investigated
chapter
associations between a concurrent measure of attachment, the
Strange
Situation,
modified
and the child narratives. The predicted association betweeii
the child's currant attachment classification to mother and/or father and their stories,
however,was not confirmed. Questionsregardingthe constructvalidity of the measure

decided
to
that by looking at the
assess
preschool
attachment
used
were raised and it was
by
instrument,
quality of early attachment, measured
a validated
more meaningful
influencing
factors
in
be
the
gained regarding
children's narrative
information might
later
functioning
in
is
be
Early
thought
to
to
related
quality of
attachment
construction.
communication and affect regulation and this chapter investigates the question of
Story
MacArthur
Stem
between
to
the
the
the
results of
associations
Battery and their attachment security assessedat twelve months of age with mother and
from
Situation
Phase
Strange
11
father
the
procedure
using
eighteen months of age with
Two.
Chapter
described
Project
Parent-Child
London
in
as
of the
In contrast to the
developmental
literature
is
the
surrounding
previous chapter there much more substantive
data
Longitudinal
infant-parent
assessIng
attachment
can
status.
attachment
of
sequelae
development
later
information
the
of attachment
on
provide valuable
The story stem play narrative technique seems to hold promise for researchers
influences
the
the
earlv
relationship
experiences
of
on
possible
to
attemptingL- investioate
It,
164
inner world of the preschool child. The battery covers a range of domains in a s,N'stematic
fashion and has been found to be a good elicitor of responses that elucidate the
representationof experiencesof family relations and relationships, conflicts and defences.
(Buchsbaum & Emde, 1990). They also offer a way to look at the child's representations
of thematic content while also addressingbehavioural style or the way in which the child
conveys these themes.
One of the original aims of the MacArthur working group developing these
narratives was to create an assessmentof attachment in preschool and school age children
did
behavioural
They
which
not rely exclusively on
observations.
wished to take into
developing
the
account
abilities of preschool children, central among them language
in
in
to
the
skills, an attempt access representationsof attachment and doing so bridge the
infant
between
described
the
attachment
assessment
and
adult
methodologies
gap
earlier
in this thesis (Bretherton, Ridgeway & Cassidy, 1990).
This study is the first to apply the recently developed MacArthur Narrative
Coding Schemeto a group of non-clinical five yearold children whoseattachmentstatus

is
known.
both
parents
at infancy with
This chapter refers to many data regarding the
infant-father
infant-mother
both
to
and
attachment.
story stems and possible associations
Some a priori predictions are made while other investigations are post hoc and
hypotheses.
there
are no a priori
exploratory and, as such,
The main qUeStionto be addressed in this chapter remains. will the behavioural
be
to
the
able
child's
predict
narrative
performance
in
infancy
attachment
of
assessment
165
Stem
Battery
MacArthur
?
That
is,
Story
does a behavioural assessmentof the
the
in
internal working model of attachment collected prior to the emergence of the verbal self
predict the nature of narrative accounts of children at age five ?
Keeping in mind recent theories regarding a communication perspective of
discussed
in
Chapter
Two,
it
is
theory
attachment
predicted that securechildren will give
higher
that
are of
more responses
quality and more open and that they will use more
themesof discipline and non-excessivepunishment and positive matemal representations.
Furthermore, it is predicted that
insecure children will give more responses of a
controlling and negative nature.
No predictions will be made regarding differences between the various groups of

insecurely attached children.
Many researchers have had difficulty
in predicting
differences between insecure-avoidant and insecure-resistant children and it has been

in
limitation
the
these
than
system, perhapswe need
that,
as
a
misses
see
rather
suggested
between
the
minimization and maximization
to take more seriously
complex relationship
individuals
instance,
for
that
most
rely
as clinical experience shows us
of affective styles
hypothesis
Therefore,
1993).
(Slade,
no a priori
on a mixture of these two strategies
be
insecure
between
differences
types
the
of attachment classification will
regarding
be
insecure
types
the
A
attachment
will
post-hoc, exploratory analysis of
made.
discussed.
conductedand
with story stem techniques often ask

Another question that researchersN\-orkin(,,,
For
the
story.
example.
which
stories
of
each
are
similar
knowled.
specifics
about
(:,
is
Z7,1
, 7c
Chapter
Some
?
different
these
questions
were
addressed
of
in
which
and
other
to cach
166
Three, however, the current chapter will attempt to address these questions
in relation
to attachment. Are there particular stories which might be better at distingui shing the
from
the secure groups ? While many of the story stems describe situations
insecure
lives
the
to
common
of five-year-old children, three of the stories, Burned Hand,
Separation and Reunion, might be said to deal with attachment
related issues. Separation
Reunion
and
stories ask the child to respond to a scenario in which both parents leave on
a short trip. The Burned Hand story asks the child to respond to how the parent vill
in
the face of injury to the child protagonist. Because of the attachment-related
react
nature of these stories, it is predicted that these stories will
be among those
distinguishing insecure from secure children.
A secondary aim of this chapter is to addressthe relatively unexplored area of the

between
father
the
to
child's
relationship
attachment
compared and contrasted with the
This
to
child's attachment
mother.
chapter undertakes to add to the information
longitudinal
the
concerning
differential effects of mother and father attachment.
Attachment researchershave frequently argued that the security of the child-mother and
is
independent
be
different
to
are
reflecting what
presumed
child-father attachment
&
Wambolt,
Steele,
Steele
&
Emde
1996;
Fonagy,
(Oppenheim,
1991).
interactive styles
In PhaseOne and Two of this longitudinal project, the Adult attachment Interview was
before
birth
first
The
Strange
Situation
the
the
the
to
child.
of
parents
results
adininistered
by
the
attachment classification of the respective parent on the
predicted
were powerfully
AA1 (Fomigoy,Steele &Steele, 1991). Based on these findings, the authors concluded that
for
his/her
independent
develops
models
major attachment relationships based
the infant
individuals.
history
Therefore,
these
hisdier
with
each
in
interactions
of
past
we might
on
167
expect to find that different aspects of the narratives of the children prove important
depending on whether we are looking at the narratives in relation to the mother-infant
infant-father
history.
The expectation is that different
the
or
attachment
attachment
be
the
narratives will
associatedto attachment in infancy for mother and for
elements of
tather.
7.2 METHOD
7.2.1 Design
The infant-parent attachment was assessedusing Ainsworth's (Ainsworth et al,
1978) Strange Situation procedure during Phase 11of this longitudinal study (see H.
Steele, 1991 and M. Steele, 1990). The Strange Situation was conducted with mother
father
The
12
the
at eighteen months of age.
months of age and with
infant was
when
111,
Project,
Phase
in
the
the
when the
most recent phaseof
story stems were administered
battery
The
five
testing
of
a
procedure
was
part
story
stem
of
age.
years
children were
has
been
This
in
hours
half
the
between
the
two
parents.
absenceof
that took
and
one and
described in detail in Chapter Three.
7.2.2 Subjects
in
Chapters Three,
described
The subjects are the same children as previously
Four and Five. To remind the reader, ninety-two children took part in the study, with
(43
The
the
of
children
age
mean
stems.
story
eleven
all
completing
actually
eighty-nme
For
59-65
(range
the
61
boys)
of
children
eighty-six
49
months).
months
was
girls and
for
however.
three
the
to
the
their
of
story
stems,
responses
of
there are complete records
difficulties
the
the
technical
sound
and
so
responses
involvin(-,
there
were
children
It
to
the
are
unavailable.
coding
system
was
possible
of
portion
tile
content
cOj,ccj-,jjjj,.,,
168
from
data
the video tapes of these children.
the
code
performance
The subjects for whom there is complete data varies according to instrument and is
below.
in
Table
7.1
presented
Table 7.1:
Infant-parent security grouped by instrument, forced classification iii
parentheses.
with Mother at 12 months
A
Strange Situation
27(4)
1 55 (0)
Story Stems
29(3)
50(0)
insecure-avoidant
secure
with Father 18 months
DI
6(5)
97
25(3)
61 (1)
0(0)
90
9(4)
1
88
24(3)
(1)
55
1
0(0)
1
41
1
82
C- insecure-resistant
D- disorganised
7.2.3 Procedure
The procedure used to gather the infant-parent classification of attachment with
has
Strange
Situation,
Ainsworth's
father
18
12
months,
at
months and with
mother at
been described in Chapter Two and more specifically elsewhere (Steele, M., 1990;
has
Story
Stem
Battery
MacArthur
the
The
Steele, H., 1991).
method of administering
Three.
in
Chapter
detail
described
been
in
also
different
four
brief
is
the
following
The
infant-parent attachment
reminder of
a
classifications.
Secure (B) - The infants who are classified as securely attached use their mother as a
her
In
the
they
frorn
to
absence
base
unfamiliar
environment.
explore
x\hich
secure
her
but
be
distressed,
positivel,,,,
on
greet
reunion
and
maly
and
their
exploration
rcdUCe
169
then return to exploration.
Insecure-avoidant
(A) - The insecure-avoidant infants explore with little reference to
the parent, shows little distress to his/her departure and appears to ignore or avoid the
his/her
parent upon
return.
Insecure-resistant
(Q-
The third pattern, insecure-resistant or insecure-ambivalent
fail
typically
to move away from mother, and so do little exploration, they are
child
highly distressed upon separation and remain difficult to settle on reunion.
Insecure-disorganised
(D) - This fourth pattern, insecure-disorganised is a recent
from
in
it
derived
clinical samples which was noted that a small number
categorisation
in
Overall
be
the
three.
these
one of
other
confidently classified
of children could not
for
handling
have
do
to
exploration and
a coherent strategy
not appear
children
inexplicable
behaviours
in
except in the context
which were
attachment, and engage odd
in
fear
the presence of the mother.
or confusion
of
7.3 RESULTS
This sectiondescribesthe associationsbetweenthe four scalesderived from the

tactor analysis of the children's responsesto the MacArthur Story Stem Battery (using
the MacArthur
Narrative Coding Scheme)
and the
attachment
father
18
The
I-'
data
the
fication
and
at
months.
months
results
at
of
with mother
classi
for
by
father.
followed
first
Four
those
sets
of
results
are
addressed
are
mother
with
170
data.
first
Presented
for
infant-mother
the
attachment
are the results utilising
presented
insecure vs secure classifications only. This is followed by a post hoc analysis using the
insecure-avoidant,
(AB- secure and Cthree
category
of
classification
system
original
finally,
insecure-resistant) and
the four-way classification which includes the
disorganised category is shown.
These analysesare followed by a more detailed look at the results*of those factors
Each
significantly
associated
attachment
proved
with
infant-mother
classification.
which
factor that proved significantly associated with infant-mother attachment is examined
battery
in
the
the
that was administered to the children.
eleven stories
with each of
Because one of the aims of this thesis is to explore the MacArthur Story Stem
Battery and the MacArthur Narrative Coding Schemeas thoroughly as possible, the story
from
Chapter
individual
by
followed
be
the
reliable variables
an analysis of
analysis will
Three with infant's attachment classification with mother at twelve months.
The second section of the results will consider the data regarding the child
for
data.
described
father
above
infant-mother
at eighteen months as
attachment to
father
insecure-resistant
Because none of the children were classified as
with
at 18
by
data
the
the
the
to
ginal
children
in
ori
1
grouping
examine
possible
not
it
is
months,
Therefore,
three-way system of categories.
insecure
from
the
vs secure
results
disorganised
the
category
three-categor-v
including
classification
a
and
classifications
for
factorwhich
by
followed
the
These
an
story
analvsis
individual
also
are
are presented.
the
variables
with
of
reliable
individual
infant's
an
analysis
and
provcd signiticant
171
father
attachment classification with
at eighteen months.
data and the story stem factors
7.3.1 Infant-mother
Independent sample t-tests were run first to determine if there were associations
between the four factors and an insecure versus a secureattachment to mother when the
12
Levene's
for
test
the equality of the variances was observed and
months.
infant was
the values for the equal and unequal groups were used accordingly. The results are
below.
Table
7.2
presented in
Table 7.2 : Means and SDS of the four story stem factors grouped by infant insecurity vs.
security with mother at twelve months.
Means (SD)
Factor
insecure
Positive maternal
representation
Note : ** p< 01
.
df
T-value
secure
1.7008 (. 26)
1.8281 (. 22)
86
2.41
1338 (. 07)
1868 (. 10)
83
2.73
5334(. 18)
.
1252(. 12)
.
86
44
.
06
.
.
5521 (.21)
.
1237 (. 11)
.
Significance
83
01**
.
008**
.
66
.
95
.
L
(two-tailed levels of significance)
The results from the above table indicate that there are two significant results at
levc] p:! 01 or better. Both the Quality/Open Response and the Discipline/Punishment
.
tactors proved to be significantly associatedwith infant-mother attachment classification.
Children who were classified as secure with mother at 12 months of age sho\,ved
to
Responses
to
the
Quality/Open
and
referred
narratives
levels
higlier
of
sioiiificaiitl\,
172
more Discipline/Punishment themes in their narratives than children classified as
insecure.
7.3.1a
Controlling
The reader will remember from Chapter Three that father's social class was found
to be significantly related to the Discipline/Punishment factor. In order to discover if
there is an interaction effect of father's social class with the infant-mother attachment
classification on this factor a two-way analysis of variance was conducted. Table 7.3
below presentsthe means, standarddeviations, F-values and levels of significance of the
three groups of father's social class with infant security to mother at 12 months for the
Discipline/Punisliment factor.
Table 7.3 : Means and SID of father's social class and security of attachment to mother
for
12
the Discipline/Punishment factor.
at
months
Means (SD)
Discipline/
punishment
1
intermediate
professional and
managerial
occupations
partly skilled
and skilled
insecure
secure
insecure
secure
insecure
secure
1193
.
(. 05)
1597
.
(. 09)
1433
.
(. 08)
1703
.
(. 10)
1152
.
(. 03)
2742
.
(. 10)
F-value
(df)
2.86
(2,82)
07
.
As can be seen in Table 7.3, there is no significant interaction effect between

father's social class and the infant's attachment classification at 12 months. Although p
07 is considered a trend, the result is still non-signif icant and, therefore, father's socIal
.
in
following
for
infant-mother
data.
be
the
the
analyses
of
controlled
class will not
17)
7.3.1b
Post hoe analyses
Giventhat there were significant associationsbetweentwo of the four factors

at
the insecure versus secure level of categorisation, the next question to be addressed
Is,
differences
in the children'sresponsesto the story stemsbe associated
the
would
with the
more specific categories of the infant-mother insecure classifications ? In order to
address this question two analyses of variance were conducted. The first, possible
between
associations
secure and the avoidant and resistant types of insecure
classifications of attachment, are presented below in Table 7.4. This analysis uses the
original three-way system of classification.
Levene's test for the homogeneity of
features
for
variance and other
parametrictesting were observedand indicatedthat for
Response,Controlling/Negative
theQuality/Open
andPositiveMaternalRepresentation
factors a parametric one-way analysisof variancewould be the appropriate test. However,
for the Discipline/Punishment scale there was a significant indication (Levene's 3.54, p
03) that homogeneity of variance was skewed and so the equivalent non-parametric
.
test, the Kruskal-Wallis one-way anova, was used (see Table 7.5).
Table 7.4 - Means and SDS of story stem factors grouped by original three-way
classifications of infant attachment with mother at 12 months.
Factor
Means (SD)
Quallt-v/openresponse
Positive matemal
f-ratio
f-prob
avoidant
(n=29)
secure
(n=50)
resistant
(n=9)
1.7261 (25)
1.8291 (.22)
1.6195 (.29)
2,85
3.55
03*
.
5-334(19)
4741 (. 17)
.
0682(. 08)
.
2.85
1.02
2,92
1.12
32
5763 (.22)
.
1-3396
(. 11)
.
1252 (. 12)
representation
Note *p< 05
.
df
( two-tailed levels of significance)
11
174
Table 7.5 -. Mean rank of the Discipline/Punishment factor grouped by three-way

twelvemonths.
at
mother
with
classification
attachment
Mean rank
Factor
avoidant
secure
resistant
38.05
47.87
30.50
Note -p levels are basedon two-tailed levels of significance
df
h-value
significance
5.0823
07
.
Table 7.4 shows that the Quality/Open Response factor indicates a significant
level (p = 03) of distinction between the three levels of attachment classification with
.
is
the
It
12
on
the
responses
more
give
children
who
secure
months.
mother at
Scheff6's
However,
factor
Quality/Open
than either the avoidant or resistant children.
difference
there
that
no
significant
was
test
stated
which
conducted
also
was
post-hoc
between any two groups at the level of p =. 05. For the Kruskal-Wallis one-way analysis
for
difference
ties
[The
7.5).
this
(Table
is
is
correcting
value
h-value
in
the
reported
&
Castellan,
(Siegel
25%
tied
less
the
are
than
observations
that
of
andmeans
negligible
9
for
be
does
factor
to
any one
significant
not prove
1988). ] The Discipline/Punishment
direction
in
the
the
children
trend
secure
of
there
a
is
category of insecurity, although
the
resistant
Punishment
than
or
themes
avoidant
either
Discipline
and
using more
children.
levels
deviations
of significance
and
Table 7.6 below gives the means, standard
four-way
by
of
secure,
classification
the
attachment
factors
infant
for the four
considered
disorganised.
and
resistant
avoidant,
n
175
Table 7.6 : Means and SDS of the four story stem factors by four-way classification of
infant securitv with mother at 12 months.
Means (SD)
Factor
Quality/Open response
Positive maternal
representation
f-prob
12
secure
(n=50)
resistant
(n=5)
disorganised
(n=7)
1.7117 (. 26)
1.8281 (. 22)
1.7111 (. 15)
1.6530 (. 34)
3,84
1.99
1352(. 07)
1868(. 10)
1030(. 09)
1540(. 04)
3,81
2.43
07
.
3,84 1 1.96
12
5580(. 22)
.
5334(. 18)
.
3830(. 10)
.
6508(. 18)
.
1455 (. 12)
1252(. 12)
0364(. 03)
.
1061 (. 08)
.
I
f-ratio
avoidant
(n=26)
Control I ing/negative
df
.
I
.
I
3,81
30
.
1.23
The above table (Table 7.6) indicates that none of the four factors prove
indicated
for
better.
Again
05
level
the
trend
there
a
is
or
of
p=.
a
significant at
Discipline/Punishment factor that suggeststhe securechildren are giving more responses
for this theme than any of the categoriesof insecure children. However, caution must be
disorganised
for
the
the
are
children
and
to
these
resistant
sizes
group
as
results
applied
Scheff6's
for
the
the
post
statistical analysis.
results
too small to ensure the reliability of
between
difference
any two groups
there
hoc test was not observed as
was no significant
05.
level
the
of p =.
at
disorganised
discovered
is
it
category
the
that
Because
recently
relatively
possible
be
the
separate
analysis
was
a
results
obscuring
might
which identifies children
from
The
the
disorganised
analysis.
excluded
the
were
children
which
undertakenin
in
below.
is
Table
7.7
for
this
presented
group
of
variance
the
analysis
way
one
result of
for
the equality of the variances was observed.
Levene's
test
Again,
176
Table 7.7 : Means and SDS of story stem factors by three-way classification of
disorizanised
12
the
months excludinpuoun.
attachment with mother at
Means (SD)
Factor
Positive maternal
representation
Note :
f-prob
secure
(n=50)
resistant
(n=5)
1.7117 (. 26)
1.8281 (. 22)
1.7111 (. 15)
2,78
2.27
1352 (. 07)
1868 (. 10)
1030 (. 09)
2,76
3.44
03
.
(. 10)
2,78
1.71
18
0364 (. 03)
.
2,76
1.72
.
I
f-ratio
avoidant
(n=26)
Control I ing/negative
df
(. 22)
5334 (. 18)
.
1455 (. 12)
1252 (. 12)
5580
3830
.
.
.
( two-tailed levels of significance)
*p< 05
.
As can be seen from Table 7.7, there is again a significant result for the
Discipline/Punishment factor once the disorganised children have been excluded forrn
in
be
Again,
the group.
exercised interpreting the results as the group size
caution must
hoc
SchefWs
is
test
was observed and
post
the
small.
extremely
resistant children
of
for
level.
05
The
different
the
result
at
p=.
two
that
significantly
are
groups
no
confirmed
longer
that
the
is
factor
Quality/Open
significance stems
and
suggests
significant
the
no
from the inclusion of the insecure-disorganisedchildren.
7.3.2 Infant-mother
data per story
In order to increase the knowledge about the specifics of each story and to
determine if there was any one story or a pattern of stories that was contributing towards
the significant
for
results
the two
factors of
Open/Positive
Response and
further analyses were conducted. The first factor, that of
Only
Chapter
Three).
(see
those
Response.
stories
story
per
calculated
was
'()Pcii
Qualitv
10
18
177
for which the internal consistency was high enough (- > 40) for this factor were used.
.
In the case of the Quality/Open Response factor, the Separation story did not meet this
it
dropped
from
34)
(the analysis. Independent sample t-tests
and so was
=.
criterion
for
Levene's
test
the equality of the varianceswas observedand the
were conductedand
for
Table
below
7.8
the
equal
and
unequal
groups
were
used accordingly.
values
gives
be
important
determining
factor.
to
this
stories
are
proving
us indication of which
in
Table 7.8 : Means and SDS of insecure vs secure attachment with mother at 12 months
by story for the Quality/Open Response factor.
Means (SD)
Story Stem
df
T-value
Significance
03*
.
16
.
56
.
02*
.
09
.
13
.
16
.
10
.
02*
.
insecure
secure
Spilled Juice
1.7646 (. 35)
1.9233 (. 32)
86
2.20
Mother's Headache
1.6082 (.29)
1.7022 (. 33)
86
1.39
Three's a Crowd
1.7617 (.41)
1.8100 (. 36)
86
Burned Hand
1.7193 (. 32)
1.8822 (. 33)
86
58
.
2.27
Lost Keys
1.7719 (. 32)
1.8922 (. 34)
86
1.68
Sweet Shop
1.6082 (.28)
1.7033 (. 30)
86
1.50
Reunion
1.7953 (.44)
1.9178 (. 36)
86
1.41
Bathroom Shelf
1.7690 (.40)
1.8922 (.29)
86
1.66
Exclusion
1.5804 (.42)
1.7744 (. 35)
86
2.33
Biscuit Tin
1.6944 (.47)
1.8922 (.28)
86
2.27
Note :*p<
05
.
02*
.
02)
Hand
(p=
Burned
03)
(p=.
that
Juice
the
Spilled
The
indicate
story
and
story
.
both
Responses
Quality/Open
the
task
the
to
and
examiner
at
more
give
secure children
indicate
Biscuit
Tin
Story
Exclusion
The
level.
a
significant
and
also
a significant
178
difference between the insecure and secure groups (both at level of significance p=. 02)
with the secure children giving more of these responses.
The results of the analysis for the second factor, Disc ipI i ne/Punishment, by story
is presentedin the table below (Table 7.9). The internal consistency for this factor across
the stories was high (- > .40) and so all stories were used. Again, independent sample
t-tests were used and Levene's test for the homogeneity of variance was observed.
Table 7.9 : Means and SDS of secure vs insecure attachment with mother at 12 months
for each story on Disc i pline/Punishment factor.
Story Stem
Means (SD)
insecure
2037
.
0833
.
1204
.
1806
.
0556
.
3148
.
0231
.
0185
.
(.25)
T-value
Significance
83
1.98
82.71
1.99
05*
.
05*
.
83
83
38
.
1.66
81.51
1.49
secure
3299 (. 31)
.
(. 16)
1633 (.20)
Mother's Headache
.
1054 (. 16)
(. 19)
Three's a Crowd
.
2721 (. 26)
(. 22)
Burned Hand
.
1156 (.20)
(. 16)
Lost Keys
.
2959 (.24)
(. 21)
Sweet Shop
.
0272 (. 09)
(.08)
Separation
.
0646 (. 18)
(. 08)
Reunion
.
2041 (. 24)
1528 (.23)
Bathroom Shelf
.
.
1190 (.20)
0602 (. 16)
Exclusion
.
.
3571 (.26)
2593 (. 26)
Biscuit Tin
.
.
05 ( two-tailed levels of significance)
Note :*p<
.
Spilled Juice
df
97
.
70
.
10
.
14
.
71
.
83
.
12
.
33
.
82.43
1.50
13
83
1.68
83
83
72.75
83
37
.
21
.
1.54
.
09
.
From the table above (Table 7.9). it can be seenthat two stories,,Spilled Juice and
for
factor
05.
level
There
this
Headache,
of
p= .
at a
is also a
Mother's
are significant
179
trend indicated for the Burned hand (p =A 0) and the Biscuit Tin Story (p = 09).
.
7.3.3
Infant-mother
data per variable
In the spirit of investigating this relatively new and unexplored measure

thoroughly, the individual variables comprising the MacArthur Narrative Coding System
be
in
detail
in
examined
will
relation to infant-mother security of attachment. Table 7.10
independent
the
the
results of
presents
sample t-tests and the content and parental
MacArthur
Narrative
Coding
Scheme.As usual, Levene's
the
variables
of
representation
test for the homogeneity of variance is observed.
Table 7.10 : Means and SDS for the content and parental representation variables
insecure
by
grouped
vs secureattachment classification with mother at 12 months (n=86).
Means (SD)
Variable
insecure
Affection
Guilt/reparation
Excludes other

Physical punishment
father
Disciplining
Disciplining
0627
.
0960
.
.
0555
.
0707
.
.
mother
Positive mother
Shame other
Verbal punishment
01
! .
(. 07)
(. 10)
(. 14)
1263 (. 17)
Physical aggression
Note :*p
1136 (. 14)
.
0379
.
(. 11)
(. 09)
1010 (. 10)
2273
13))
1338
1')
0884 (. 10)
.
1.2197 (. 1
-33)
** p :! 001
.
t-value
83
05
.
24
.
63
.
30
.
1.88
96
.
80
.
53
.
secure
1010 (. 14)
Dishonesty
df
1150 (. 14)
1076 (. 10)
83
(. 09)
83
0555 (. 11)
.
1577 (. 15)
.
83
1262 (. 13)
83
00
.
99
.
(. 08)
83
63
.
53
1039 (. 15)
83
1.13
26
.
1150 (. 11)
83
58
56
.
0501
.
.
0694
.
.
.
3173
.
1354
.
1299
.
1.2968
83
(. 18)
83
21.66
(. 13)
83
06
.
(. 14)
82.98
1.45
(. 19)
182.76
12.19
76
06
.
009**
.
95
.
15
.
03
.
180
Of the content and parental representation variables. Table 7.10 shows that two
infant's
12
the
to
security
of
attachment
with
mother
at
months.
related
are significantly
These are 'disciplining mother' (p = 009) and 'verbal punishment' (p = 03), not
.
.
included
in
factor
that
these
two
the
variables are
surprising given
for
(above).
A
trend
appears the variable 'excludes other'. also
which proved significant
included in the Discipline/Punishment factor.
Table 7.11 presents the results of the independent sample t-tests for the
for
homogeneity
Levene's
As
NINCS.
the
test
the
of
usual,
performance variables of
variance is observed.
181
Table 7.11 : Means and SDS of performance variables grouped by insecure vs secure
bv
12
(n=89).
months
at
attachment with mother
Means (SD)
Variable
df
t-value
06
.
43
.
01*
.
06
.
50
.
77
.
18
.
04*
.
18
.
05*
.
insecure
secure
Role of parent
1.6914 (. 47)
1.8691 (. 42)
86
1.86
1.0335 (. 15)
1.0600 (. 15)
86
79
1.8062 (. 34)
1.9582 (. 15)
48.43
.
2.50
2.6005
2.7909
65.20
1.90
Involvement of examiner
1.4665 (. 30)
1.4255 (. 26)
86
2.1699
2.2127
(. 66)
86
Denial
2600 (14)
.
1.7800 (. 39)
86
3086 (. 19)
.
1.6100 (. 37)
67
.
29
.
1.33
86
2.06
Narrative coherence
2.6053
2.7564
(. 51)
86
1.35
52.79
2.01
86
1.23
(. 52)
(. 72)
(. 52)
(. 38)
0709 (. 13)
.
.
5339 (. 25)
4681 (. 24)
.
.
0594 (. 05)
0534 (. 06)
.
.
0285 (. 03)
0199 (. 04)
.
.
0442 (. 05)
0431 (. 07)
.
.
6060 (. 30) 1 5861 (. 26)
.
.
1643 (. 26)
Control
Joy
Anger
Distress
Concern
Anxiety
Note :*p
! .05
** p ! 01
.
86
86
86
86
47
.
98
.
08
.
34
LIL
64
.
33
.
93
.
1.73
--
based
on two-tailed tests
p values are
Narrative
MacArthur
from
the
As can be seen,two of the performance variables
to
Response
Adaptive
related
significantly
are
code,
System
Coding
and the additional
is
'Directness
12
significant
style'
of
performance
months.
at
mother-child attachment
direct
style
performance
01
more
a
the
showing
level
children
secure
of p= . with
at a
indicating
05
level
that
the
is
'Control'
of
p=
at
a
also significant
than insecure children.
.
'Adaptiveness
displaying
than
of
children.
secure
control
more
insecure children are
higher
04
level
the
showing
children
secure
%vith
p=
of
.
a
at
significant
also
is
response
.
for
be
direction
is
to
There
trend
the
children
secure
in
a
also
le-v,
els of adaptiveness.
182
more responsive to the examiner (p = .06) and for the secure children to utilize at least
doll
(p
06).
one parent
=.
7.3.4 Infant-father
data and the story stem factors
Independent sample t-tests were run using the infant-father data and the four story
factors
deten-nine
if
to
their might be significant associationsas there was with the
stem
infant-mother data. Levene's test for the equality of the variances was observed and the
for
the equal and unequal groups were used accordingly. The results are presented
values
in Table 7.12 below.
Table 7.12 : Means and SDS of the four story stem factors grouped by infant insecurity
father
at eip-hteenmonths.
vs. security with
df
Factor
insecure
Control Iing/negative
Positive maternal
representation
Note : *p : 05
.
T-value
secure
1.6881 (.30)
1.8044 (.23)
81
1.91
1660 (. 10)
1659 (. 09)
78
01
.
39
.
31
.
5545 (. 25)
.
1344 (. 11)
.
Significance
.
5359 (. 17)
.
1252 (. 12)
.
81
78
05*
.
99
.
70
.
76
.
As Table 7.12 presents,only one of the four factors proved to be significant at the
Response.
This
it
is
Quality/Open
level,
05
that
the children
that
result reveals
of
p :! .
father
be
18
'udged
to
their
to
attached
secureINI,
at
months of age who are able
wlio \vere
5
to
the
lit(-)her
more
open
responses
and
story stems at age years than the
quality
to olve
181
children classified as insecure with their father at 18 months of age.
7.3.4a Post hoc analyses
Again, given that there were significant associations between

one of the four
factors at the insecure versus secure level of categorisation, the
next question to be
addressed is, would the differences in the children's responses to the story stems be
associatedwith the more specific categoriesof the infant-father insecureclassifications

As there are no children in the sample who were classified resistant with father
at
it
is
eighteen month
not possible to examine the original three-way classification of
Ainsworth's. Therefore, the first analysis of variance considers possible differences or
between
variations
secure, avoidant and disorganised types of insecure classifications
is
below
in Table 7.13. Levene's test for the homogeneity
of attachment and presented
of variance was observed and indicated a parametric one-way analysis of variance would
be the appropriate test.
Table 7.13 : Means and SDS of the four story stem factors grouped by three
father
infant
18
attachment
securit-v
classifications of
with
at
months.
Factor
Disc ipl ine/pun ishment
Positive maternal
_representation
df
f-ratio
f-prob
avoidant
(n=24)
secure
(n=55)
disorganised
(n=4)
1.6776 (. 30)
1.8157 (.22)
1.6831 (28)
2,80
2.73
07
.
1618(. 10)
1720(. 09)
1061 (10)
2,77
84
.
43
.
10
90
.
03
.
96
.
5571 (. 25)
.
5357 (. 17)
.
5277(. 08)
.
2,80
1-302(. 11)
1279(. 12)
1136(. 09)
2,77
184
Table 7.13 shows that there are no significant differences. There is a trend in the
direction of secure children giving more Quality/Open responsesthan
either the avoidant
or disorganised children, however, caution in accepting these results is urged as group
for
disorganised
the
numbers
group is so small (n=4) that statistical analysis is
for
this group. Given this, it was decided to do a further analysis excluding
meaningless
the disorganised children from consideration. This is presented below (Table 7.14).
Table 7.14 presentsthe independentt-tests for the avoidant and securegroups with father
excluding those children classified as disorganised. Levene's test for the homogeneity
of variance was observed and the equal and unequal groups were used as appropriate.
Table 7.14 : Means and SDS of four story stem factors grouped by avoidant and secure
father
attachment with
at 18 months (excludes disomanised children).
Means (SD)
Positive maternal
representation
Note :*p:!
.05
T-value
05*
.
avoidant
(n=24)
secure
(n= 55)
1.6776 (. 30)
1.8157 (.22)
77
2.24
1618(. 10)
1720(. 09)
74
41
.
5571 (.25)
.
5357 (. 17)
.
77
1302(. 11)
1279(. 12)
74
43
.
07
.
df
68
.
67
.
94
.
(two-tailed tests of significance)
It can be seen from Table 7: 14 that, once the disorganisedchildren have been
for
factor
Quality/Open
is
from
the
the
the
sample, result
onceagain significant
CXCIUded
(p
O)
185
7.3.5 Infant-father
data per story
As was done with the infant-mother data above, further analYseswere conducted
in order to determine if there was any one story or a pattern of stories that was
for
factor.
Quality/open
The
Response
the
towards
the
significant
result
contributing
Quality/Open Response factor was calculated per story
(see Chapter Three).
Independent sample t-tests were conducted and Levene's test for the equality of the
variances was observed and the values for the equal and unequal groups were used
Table
7.15 below gives us indication of which stories are proving to be
accordingly.
in
determining
factor.
The
Separati'on Story Iis excluded from the analysiis
this
important
for
internal
the
as
rating
consistency was below .40.
Table 7.15 : Means and SDS of secure vs insecure attachment with father at 18 months
for each story on Quality/Open Response factor.
Means (SD)
Story Stem
df
T-value
81
.
19
.
38
.
42
.
84
.
70
.
insecure
secure
Spilled Juice
1.7978 (. 33)
1.8649 (. 35)
81
Mother's Headache
1.6556 (. 33)
1.6705 (. 32)
81
Three's a Crowd
1.8067 (.40)
1.7711 (. 38)
81
Burned Hand
1.7333 (. 34)
1.8295 (. 33)
81
1.19
23
.
Lost Keys
1.7867 (. 30)
1.8630 (. 35)
81
94
.
34
.
Sweet Shop
1.6333 (. 32)
1.6743 (.29)
81
56
.
57
.
Reunion
1.7111 (.47)
1.9224 (. 36)
81
2.20
Bathroom Shelf
1.7044 (.44)
1.8870 (.29)
81
2.20
Exclusion
1.4533(. 48)
1.7989 (. 32)
1.7267 (.40)
Biscuittin
Note :*p<
05
.
1.8410(. "6
)
levels
(t\N,
of sionificance)
o-tailed
Z--
33.42
1- -T
81
3.28
I. 20
03*
.
03*
.
002**
.
.2-3
186
Three stories proved able to differentiate the insecure from the secure children.
Reunion story, Bathroom Shelf and Exclusion were all significant. The results show that,
in all three stories, secure children have stories that are more open and of higher qual't-Y
than the insecurechildren.
7.3.6 Father-data per variable
The individual variables comprising the MacArthur Narrative Coding Systemwill

be examined in detail in relation to infant-father security of attachment. Table 7.16
for
independent
t-tests
the content and parental
the
the
sample
results of
presents
insecure
by
NINCS
the
versus secureattachmentwith
grouped
representationvariables of
father at 18 months. As usual, Levene's test for the homogeneity of variance is observed.
187
Table 7.16 : Means, SIDS and levels of significance for the content and parental
insecure
by
vs secureattachment classification with father at 18
months (n=86).
Variable
Means (SD)
insecure
Affection
Dishonesty
1146 (. 14)
Guilt/reparation
Excludes other
Physical aggression
Physical punishment
Disciplining father
Disciplining mother
Positive mother
.
0474
.
0359
.
1265
.
1502
.
0843
.
0672
.
0949
.
(. 09)
(. 09)
(. 12)
(. 18)
(. 13)
(. 09)
(. 10)
2846 (. 16)
.
1304 (. 11)
.
1621 (. 15)
Shame other
Verbal punishment
Note :*p
1383 (. 13)
05
! .
.
1 2609
.
(. 15)
df
t-value
1116 (. 14)
78
76
44
.
(. 11)
78
(. 08)
78
(. 11)
78
(. 16)
78
(. 13)
78
(. 08)
78
28
.
81
.
1.13
(. 15)
78
1.08
(. 11)
78
77
(. 17)
78
(. 13)
78
(. 11)
32.49
secure
.
0925
.
0415
.
0681
.
1372
.
1196
.
0559
.
1037
.
1164
.
2855
.
1388
.
0861
.
2663
.
(. 18)
178
.
73
.
27
.
1.16
.
02
.
26
.
2.16
1 13
.
46
.
78
.
.25
78
.
42
.
26
.
28
.
44
.
98
.
79
.
03*
.
1 90
.
Of the content and parental representation variables, Table 7.16 shows that only
father
18
infant's
to
the
at
months.
of
attachment
with
security
one is significantly related
This is 'shameother' (p = 03). The table below (Table 7.17) presentsthe results for the
.
father-child
in
to
attachment.
pertormancevariables relation
188
Table 7.17 : Means and level of significance for the performance variables by insecure
father
18
(n=89).
to
at
months
vs secure attachment
Means (SD)
Variable
df
t-value
26
.
89
.
14
.
06
.
78
.
80
.
01*
.
006*
.
26
.
20
.
45
.
88
.
42
.
65
.
82
.
insecure
secure
Role of parent
1.6909 (. 54)
1.8292 (. 40)
35.92
1.14
1.0436 (. 15)
1.0486 (. 16)
81
13
1.8073 (. 34)
1.9201 (. 23)
34.24
.
1.51
2.5236
2.7759
31.43
1.93
involvement of examiner
1.4436 (. 28)
1.4248 (. 2 8)
81
investment in performance
2.1564
(. 86)
2.2038
(. 61)
35.07
2182
.
(. 12)
3150
.
(. 18)
81
28
.
25
.
2.43
Denial
(. 60)
(. 35)
1.5345 (. 34)
1.7931 (. 40)
81
2.81
Narrative coherence
2.5927
2.7367
81
1.12
(. 18)
81
1.27
(. 26)
81
75
(. 06)
81
(. 03)
81
(. 05)
81
(. 28)
81
1
(. 56)
1491 (. 24)
Control
.
4703
.
0606
.
0303
.
0485
.
6048
1.
Joy
Anger
Distress
Concern
Anxiety
Note :*p
01
: .
** p :! 001
.
(. 23)
(. 05)
(. 04)
(. 09)
(. 29)
0878
.
5162
.
0585
.
0225
.
0413
.
5893
1.
(. 52)
.
.
15
80
.
45
.
23
1
Narrative
from
MacArthur
the
be
As can
seen,one of the perfon-nancevariables
Coding System is significantly related to father-child attachment at 18 months. *Denial'
higher
levels
denial.
01
level
the
of
showing
children
secure
with
of
p
=.
a
at
is significant
level
006
highly
is
the
secure
of
p=
with
a
at
significant
also
'Adaptiveness Of response'
.
in
direction
for
is
There
the
levels
trend
higher
also
a
adaptiveness.
of
showing
children
06).
(p
be
to
the
=
to
examiner
responsive
more
children
seCLIre
.
189
7.4 DISCUSSION
The aim of this chapter was to investigate possible associations between a

behavioural measureof attachment in infancy and a semi-structured narrative task at five
has
been
As
in
suggested earlier this thesis, attachment theory is of great
years of age.
importance when discussing the link between the parent-infant relationship and the
developing
in
the world outside of that relationship.
understanding and skills
child's
Early attachment theory proposes that an individual's early attachment history becomes
transformed or integrated into internal representationsand that theserepresentationshave
for
implications
behavioural
The
experiences.
question of whether a
predictable
infancy
could predict the outcome of the child's responses
assessmentof attachment in
to a narrative task four years later has been answered. The results described above are
be
does
to
that
there
appear
significant
most promising and exciting and confirm
infant-father
both
the
the
relationship
within
and
early
infant-mother
reflections of
narratives of preschool children.
7.4.1
Infant-mother
data
When considering the infant's attachmentclassification to mother measuredat

12 months of age by Ainsworth's Strange Situation in binary form, that is insecure or
from
be
derived
factors
four
the
to
story
stems
proved
the
significant
as
two
of
SeCLire,
factor
Quality'Open
Response
the
the
To
the
assess
coherence
reader,
remind
predicted.
directness
the
to
the
examiner,
of
performance
style
and
responsivity
of the narrative,
190
The
factor
includes
the
conflict.
such
child's understanding of
items as physical and verbal punishment, representations of a disciplining mother and
father,, shaming and excluding the other.
There is a strong association for secure children to give more responsesthat are
be
higher
judged
insecure.
There
to
than
of
a
quality
and
more open
children
considered
for
discipline
themes
to
of
and
securechildren use more
was an even stronger association
in
insecure
These
their
than
children.
resultswere true even after
narratives
punishment
the possible interaction effect of father's social class was taken into account.
In order to explore the nature of these effects in more detail and to see if
differences could be distinguished between the separate groups of insecure children
further post hoc analyses were undertaken. Mary Ainsworth's original three category
insecuredeemed
This
first
those
takes
children
system
examined.
system was
disorgamsed and forces them into one of the three original categories. One significant
insecure-avoidant
the
from
and
this
of
investigation;
exploratory
obtained
was
result
insecure-resistant groups, it is the resistant children who were least likely to tell stories
judged to be open and of high quality. When an analysis of the sample was conducted
utilising
insecureincluding
is,
the
that
the four way system of categorisation,
In
the
the
that
considering
disorganised children, it appears
are
non-significant.
results
for
it
be
that
four-way
for
group
sizes
out
pointed
must
system
the
classification
results
I
too
small
groups
were,
unfortunately,
the insecure-resistantand insecure-disorganised
to reach reliable conclusions.
191
The insecure-disorganised classification is one which has been added most
recently to observational rating systemsand remains in need of further empirical research
its
function.
Because
it
is
nature
and
regarding
possible that the insecure-disorganised
children were obscuring/skewing the results an analysis was made which excluded them
from consideration. The result was that the Quality/Open Response factor remains
level
the
significant only at
of a trend. We may speculatefrom looking at the mean scores
that it is the children who are forced-resistant that seem to be having trouble responding
high
quality and open manner. Regarding the Discipline/Punishment factor the story
in a
is reversed.The result is significant with the securechildren remalning most likely to use
themes of discipline and punishment in their stories. Insecure-resistant children are the
least likely to give responsesthat include themes of discipline and punishment. Again
four-way
is
three-way
the
and
categories of
caution urged when interpreting results of
insecure-disorganised
for
insecure-resistant
the
the
and
are
group sizes
classification as
small.
Why were there not significant results for the Controlling/Negative and Positive
Maternal Representation factors ? It was predicted that the insecure children would give
that
than
children
and
secure
secure
nature
and
controlling
of
a
negative
more responses
The
maternal
representations.
positive
children would give more responsescontaining
Control Iing/Negative factor assesses responses for themes of physical aggression,
Positive
the
the
performance, control and negative atypical response;
investment in
for
factor,
Representation
Maternal
affection and positive maternal representations.When
it
insecure
the
control
was significant and vas
the individual variables were examined.
Hovvever,
i
li
ikelv
than
the
this
more
secure
sti*ate(-,
N,
group.
it
is
using
were
who
children
192
that the main reason the Controlling/Negative factor failed to provide results is due to the
non-clinical nature of the sample.
A secondary aim of this chapter was to investigate the question of whether

be
better
particular stories might
at distinguishing the insecure from the secure with
?
It
mother groups
was predicted that three stories dealing with attachmentrelated
issues, Burned Hand, Separation and Reunion would do so but this was not the case.
From the analyses it appears that the Spilled Juice story, Burned Hand, Exclusion and
Biscuit Tin were the stories that best predicted the child's security of attachment to
However,
the
thinking
the
child's response.
about
quality and opennessof
mother when
the Reunion story did indicate differences in the insecure and secure children for
just
infancy.
father
in
Are
the
that
the
these
to
ones
address
child
stories
at
attachment
the right level of interest and anxiety for this age group ? As an administrator of many
in
Keys
for
Lost
to
the
the
one arousing
most anxiety the
seemed
example,
of the stems
for
be
hypothesised
this
the
It
that
age
emotional context of
children
of
could
children.
this story proved to be too anxiety provoking regardless of attachment classification.
If one consideredthemesof discipline and punishmentit was the Spilled Juice

Burned
Hand
Biscuit
Tin
(possibly
Headache
Mother's
and
as well) which
story and
identifying
in
be
those
to
children
who
were
securely
attached
to
significant
proved
in
infancy.
mother
In terms of the individual variables, disciplining mother, verbal punishment.
193
directness of performance style, adaptiveness of response and control proved to the
variables associated to security of mother in infancy. Of these adaptiveness of response
designed
to assessthe overall and so, in some ways corresponds to the quality/open
was
factor.
7.4.2 Infant-father
data
Turning to the infant-father relationship whereastwo of the four factors derived

from the story stems were associatedwith infant-mother attachment only one of the four
factors is associatedwith the attachment to father measuredat eighteen months. Children
father
in
infancy
higher
to
attached
give
quality and more
who were assessedas securely
five.
This
task
the
remained true when the insecureat age
narrative
open responseson
disorgamsed children were removed from the analysis.
When investigating the question of whether particular stories might be better at
distinguishing the insecure from the securewith father groups on the Quality/Open
Responsefactor it was the Reunion Story, Bathroom Shelf and Exclusion that were the
Story
the
Reunion
to
The
did
those
secure
predicted
result
in
of
one
was
that
so.
stoi-ies
is
higher
that
being
what was
quality and more open responsesand
able to give
children
factor
in
Separation
for
internal
the
this
Unfortunately,
tOLInd.
the
story was
consistency
between
The
Exclusion
the
hioh
association
to
even yield reliable statistical results.
not
high
factor
this
on
with
secure
children
particularly
was
attachment
of
security
and
story
higher
responses.
I
vely
qualitati
and
open
to
more
c
()ix,
able
IL-
From a psychoanalytic
fundamental
Oedipal
task
taps
regarding
a
clevelopmental
this
into
storv
perspective.
194
development and triangular relationships. Are those children -,
vho are deemed secure in
their relationship with father more able to understand and address this intra psychic
direct
conflict in a
and coherentmannerthan thosejudged insecurewith father in infancv
7
7.4.3
Differences in the infant-mother
and infant-father
findings
The quality and openness of the children's narrative response was found to be
both
infant-mother
the
significantly associated with
and infant-father attachment
factor
It
that
this
appears
captures elements of the child's ability to
classification.
five
is
that
related to their quality of their relationship with
construct a narrative at age
the parents in infancy. Rather surprisingly, the discipline and punishment themes so
infant's
found
the
to
mother
at
one
classification
with
year,
were not
strongly associated
to be related to the infant's attachment to father. This could be suggestive that the
influence
internalisation
has
the
on
a stronger
of values
primary relationship with mother
discipline.
to
related
Oppenheim at al (1996) propose a transactional theory which might go some way

differences
between
father
there
the
are
mother
and
towards explaining why
results. In
between
made
children's narratives with mother-child and
their study comparisons were
father-child co-constrLictednarratives. Although their study did not involve attachment
found
data,
that there was a pattern of separatecontributions from children's
they
history
195
fathers
their
to the coherence of their independent
their
mother and
interactions with
narratives and to the number of prosocial themes they constructed. They suggestedthat
the qualities of the parent and child co-constructed narratives are 'emergent properties
between
parents and children and cannot be reduced to
of reciprocal exchanges
in
one or the other'. According to this view, these emergent
emanating or residing
by
children and carried over to new contexts such as
properties are internalised
(Oppenheim
is
1996).
This
et al,
independent narrative construction
similar to objectdescribed
thinking
earlier in this thesis whereby aspectsof the
relations and attachment
infant-father
intemalised
infant-mother
infant
by
the
are
and
relationship
and
separate
transformed into internal representations that guide behaviour.
In conclusion, an exciting aspect of the findings in this chapter is that the story
be
tapping into an aspect of
to
stems seem
the early infant-parent
attachment
five.
What
in
is
the
quite
emerges
at
age
that
narratives
child
reflected
relationship
had
have
securerelationships with
strongly is that there is something about children who
both parents in infancy that enable them to give higher quality responses. The next
internal
influence
the
of
models
working
the
own
parents'
of
chapter investigates
the
child narratives.
on
prenatally
attachment, assessed
196
CHAPTER EIGHT
ASSOCIATIONS
AND THE PARENTS' ATTACHMENT
STORY STEM BATTERY

CLASSIFICATION
197
8.1 INTRODUCTION
This chapter presents the data regarding the assessmentsof the parent's adult
by
as
assessed the Adult Attachment Interview (Main et al,
1985) before the birth of the target child. The Adult Attachment Interview has been
described in Chapter One in some detail. This instrument measuresattachment securitY
by using a semi-structured interview about the parent's perceived experiencesabout their
data
in
differs
from
in
The
this
that
chapter
attachment
reported
reported the
childhood.
dyadic
Chapters
6
7
the
two previous chapters;
and were concerned with assessmentsof
from
behaviour.
The
Adult
Attachment
between
parent and child assessed
relationship
Interview assesses only the adult subject's current state of mind with respect to
history.
attachment and relationship
While the Strange Situation and the Adult
for
(either
described
be
both
Interview
Attachment
offspring's
assessing
as a means
could
infant or adult offspring) representations of his/her relationship with parents (Benoit,
Parker & Zeanah,, 1997), the Adult Attachment Interview does not measure actual
but
the
to
parents'
measure
attempts
relationship
experiences of a parent-child
being
the
the
through
semianalysis
of
their
parented
of
experiences
of
representations
link
between
the
This
the
interview.
parents'
a
of
possibility
explores
chapter
structured
Interview,
by
Adult
Attachment
the
internal working models of attachment, as measured
five
the
years.
befi)re the birth of the target child and the narrativesof
children at age
,
history
for
their own child's
their
Parents' representations of
own attachment
highly
important
be
Steele
(Fonag
&
Steele.
to
development
acknowledged
is
social
.
literature
has
However.
1994).
the
also generally conceptualised adult
1991, Nlain,
198
attachment as a somewhat more indirect and distant factor that is mediated by such
factors as parental responsiveness,family harmony and parent-child
attachment quality.
Similar to the AAL) the story completion task provokes

children to think about the
distressing
possibility of
events that may have happenedto them in real life (for example,
injury
an
or separation)and assesseshow the child deals with such a situation. The
presentation of story stems may activate the attachment system in ways not dissimilar to
AAI questions such as 'what did you do when you were hurt as a child T
From a communication perspective on attachment representations, AAI

individual's
classifications appraise an
ability to consider and discuss attachment issues
with emotional openness and coherence (Bretherton, 1995c). The MacArthur story
discuss
to
stems are asking children consider and
emotionally charged issues, including
those related to attachment. For example, the Separation and Reunion stories, as well as
Burned Hand, relate directly to the attachment construct. Separation asks the child to
happens
leave
the
tell
after
parents
on a short trip and Reunion asks them
what
show and
to say what happens immediately upon their return. In the Burned Hand story a child is
told not to touch the hot stove but does so and injures their hand. As well as thematic
here
described
in
(and
Chapter
Three
in
the
the
coding
scheme
used
stories,
content of
detail) also attempts to assessthe quality of the child's ability to consider and discuss
it
is
Quality/Open
factor
Response
In
issues.
the
these
particular,
scale derived from the
factor anal'sis in Chapter Three that refers to this aspect. The Quality/Open Response
and/or assessingnarrative coherence,responsiv itv
I'actor is comprised of scalesmeasurin-(2
directness
of performance style, child's understanding of the confllct5
to the examiner.
199
joy and use of parent dolls.
In view of theseideas,it is predictedthat the child's responseson the story stems

for the Quality/Open Response factor will differentiate those parent's classified as
insecure versus secure before the birth of the target child and that those children whose
parents' were classified as secure will produce stories of a higher quality.
The
Controlling/Negative factor addressescombined elements of negative themes such as

behaviour
high
physical aggression and negative atypical response and controlling
and
investment in the performance in the child responses. In view of the fact that these
themes and behaviours appear to be counter to the ability to consider and discuss
it
is
be
in
issues
there
that
a negative
predicted
will
an open manner,
emotional
factor
I
ing/Negati
Control
between
the
and parent security.
ve
connection
The AAI assessesparents' perceptions of their experiencesof being parented and
be
by
is
to
reflect upon and objective and nona secureadult characterised an ability
defensive in their attitude to relationships. Thesesecure-autonomousparentsare also
As
implications
developmental
experiences.
of affective and relationship
aware of the
because
disciplining
their
child
such, they may use more authoritative means of
demands
for
firm
disciplining
mature
rules,
are associatedwith
authoritative methods of
into
take
to
the
behaviour, reasoning and negotiating with
account their point of
child
view and encouraging
attachment.
1997).
(Grusec,
autonomy
Thus it is predicted that the
factor will be positively associated to parents' quality of
200
8.2
METHOD
8.2.1
Design and Subjects
The design of this study follows the model tested in the previous two
chapters
concerned with attachment data. The participants discussed in this chapter are a part of
the London Parent-Child Project sample, which has been referred to throughout this text
is
below.
and surnmarised
Mothers
The original sample (n = 100) of mothers were first-time mothers, well-educated,
Their
white and predominantly middle-class.
median age was 31 years (range of 22 - 42).
Seventy-five of the women were born in England, 10 in Scotland or Ireland while 15
born
but
Ireland,
UK
Western
the
were
outside
or
areas primarily
and Anglo-Saxon.
Eighty-two of the woman were married to the expectant father at the time of recruitment
or married subsequently.
Twenty-one of the woman represented social class I
(professional and managerial); 65 social class 11(intermediate occupations); 9 social class

III (skilled occupations), and 5 social class IV (partly skilled occupations) according to
the criteria of the UK government Office of Population Censusesand Surveys (1980).
(See also Steele, H., 1991 & Steele, M., 1990).
Fathers
Like their fernale partners, the sample of expectant fathers (n= 100) constituted a
hi(Ihly homogeneous educated,middle and upper-middle class group. Their median age
201
33
fathers
(range
Eighty-seven
20-57).
from
England, five
the
of
was
expectant
were
from
Scotland
Ireland
bom
and eight were
or
were
outside the UK, though none of these
being
different
countries
markedly
cultures. Twenty-five of the men represented social
I
(professional
and managerial); 57 social class II (intermediate occupations); 16
class
III
(skilled occupations); and 2 social class IV (partly skilled occupations),
social class
according to the criteria of the UK Government Office of Population Censuses and
Surveys (1980). (See Steele, H., 1991 & Steele, M., 1990).
The loss of parental participants since the initial prenatal assessmentamounted

to 14 caseswhich were evenly distributed in terms of maternal AAl classification (five
insecure, nine secure), but somewhat skewed toward insecurity in terms of paternal AAI
identifiable
insecure,
being
The
(seven
of
six
secure).
single
greatest
cause
attrition
status
the fact that the families had moved outside of the UK.
Children
As previously described, ninety-two of the original families agreed to participate
(43
boys)
49
The
Project.
III
the
Phase
the
girls and
children
was
mean age of
of
in this,
61 months (range 59-65 months). Attrition since the 12 month visit was 6 children (or 7
%) and since the 18 month visit was 10 children (or 10 %). One child, whose parents
but
joined
for
12
18
then
the
and
month visits
participated pre-natally, was unavailable
Eighty-nine
for
5
the year assessment.
of these returning children completed
the cohort
the eleven story stems.
202
8.2.3 Procedure
This chapter combines data from both the first and the most recent phasesof the
London Parent-Child Project. During Phase I of this longitudinal project, which was
during the last trimester of the target child's pregnancy, each parent completed an Adult
Attachment Interview. Ninety-eight of the couples were interviewed in their home and
two in the laboratory. The interviews were conducted concurrently in different rooms.
(also reported elsewhere in detail, Steele, M, 1990 & Steele, H, 1991). At age five, the
including
further
the
to
their
assessments
complete
parents returned
children and
Macarthur Story Stem Battery described in detail in Chapter Three.
The following is a brief reminder of the Adult Attachment classifications (seealso

Chapter One).
has
(F)
of
mind
characterised
a
state
Secure-autonomous
the
adult
autonomous
secure
influential
high
by the
on subsequent
regard given to attachment relationships as
development. They present themselves as self-reliant, objective and nondefensive and appearto have come to terms with past experiencesthus permitting
balanced
view of relationships.
a
deny
to
tend
experiences
(D)
negative
adults
Insecure-dismissing
insecure-dismissing
These
implications.
developmental
dismiss
adults
also
their
to
and emotions or
feelings
difficulty
have
the
the
little
associated
with
re-evoking
and
remember
to
recall.
they
able
are
experiences
20")
Insecure-preoccupied (E) - The insecure-preoccupied classification suggests a state of
They
indicative
their
preoccupation
continuing
with
parents.
of
appear
mind
incoherent
influence
their
and
unobjective
about
relationshipsand
confused,
over
them. Their anger over past and present experiencesdoes not seemto be resolved
be
in
to
they
enmeshed their early relationships and unable to proceed
appear
and
beyond them.
Unresolved (U) - These parents appear to be unresolved in their narratives regarding

their experiences of either loss or trauma. The transcripts of these subjects are
disorganised
by
irrational
the
thoughts,
of
recollections
or
generally
characterised
loss or trauma.
As stated, the assessmentof the Adult Attachment Interview was carried out
be
AAI
have
to
the
birth
stable within
before the
also shown
of the child, and studies
1994).
&
Parker,
(Benoit
individuals over time and even across generations
8.3 RESULTS
This segment describesthe associationsbetween the parent's attachment security

four
factors
derived
birth
the
target
before
the
AAl
by
the
with
child
of
the
as measured
in
the
been
has
data.
As
the
results section
chapters,
previous
pattern
trorn the story stem
data
ing
describi
then
to
to
the
and
into
mother
1
pertaining
divided
two
be
areas
main
will
between
differences
the
the
One
the
possible
of
Section
results
father.
xv-illaddress
birth
before
the
the
the
of
child
and
assessed
Is
classifications
attachment
adult
mother
204
responses that child gave on the narrative story task at age five years. Section two will
present the results for the analysis of father's adult attachment classification and the four
story stem factors.
8.3.1 Mother data
In order to detennineif there is a difference betweenthe

children's narrativesat
five
age
and the parental attachment classification (measured by the Adult Attachment
Interview) an independent samples t-test was conducted. Levene's test for the
equality
of variance was observed and the values for the equal and unequal groups used as
appropriate. The results are presented in Table 8.1 below.
Table 8.1: Means and SDS of the four story stem factors with mother's AAl attachment
(insecure
classification
vs secure) before birth of child subiect.
Factor
Means (SD)
Positive materrial
representation
Note: * p< 05
.
Table
8.1
df
T-value
insecure
(n = 36)
secure
(n = 53)
1.7357 (.25)
1.8034 (.24)
87
1.25
1377 (.07)
1839 (. 10)
83
2.36
5703 (.21)
.
5230 (. 18)
.
87
1.12
1247 (. 10)
1239 (. 12)
84
03
.
.
I
.
I
Significance
.21
02*
.
26
.
97
.
-1
aboNc shows
that
there
is
one
significant
result.
The
Disciplme, PLinishment factor is associated with mother's attachment classification as
205
measured by the Adult Attachment Interview. Children whose mothers were classified
as secure-autonomous before the birth used more themes of discipline and punishment
than those whose parents were classified insecure.
8.3.1a Controlling for father's social class
As father's social class was found to have a

main effect related to the
Discipline/Punishment factor (see Chapter Four) a two-way
analysis of variance was
determine
if there is a significant interaction effect of father's
to
conducted
social class
with mother's adult attachment classification. Table 8.2 below presents the means and
deviations
standard
of the three groups of father's social class with mother's attachment
during
for
the Discipline/Punishment factor.
measured
pregnancy
Table 8.2 : Means and SD of father's social class and mother's AAI attachment
factor.
the
classification with
Means (SD)
Discipline/
punishment
intermediate
professional and
managerial
occupations
insecure
(n= 9)
secure
(n= 13)
insecure
(n= 19)
secure
(n= 28)
insecure
(n= 6)
secure
(n= 9)
1195
.
(. 07)
1678
.
(. 08)
1380
.
(. 07)
1710
.
(. 10)
1641
.
(. 09)
2593
.
(. 11)
F-value
(di)
61
.
(2,83))
54
.
partly skilled
and skilled
As can be seenfrom Table 8.22.there is so significant interaction effect between

I'ather"ssocial class and mother's adult attachment classification and, therefore. father's
for
in
be
following
the
controlled
analysesof the mother attachment
social class will not
206
data.
8.3.2 Post hoc analyses
Given that there was a significant associationbetweenone the factors the

of
at
insecure versus secure level of categorisation (Discipline/Punishment; T-value = 2.36.
it
decided
02),
to conduct a post hoc analysis to explore if it might be possible to
p=.
was
distinguish differences between the more specific categories of insecurity. In
order to
addressthis question, two analysesof variance were conducted. The results of the first
in
Table 8.3. This table presents the possible differences between the
are presented
secure and the insecure-dismissing and the insecure-preoccupied classifications of
These
include
mother.
groupings
mothers who were given a rating of unresolved but
were also given a second or 'forced' rating of one of the three original categories.
Levene's test was again observed.
Table 8.3 : Means and SDS of story stem factors grouped by three-way classifications of
interview.
mother's adult attachment
Means (SD)
Factor
f-prob
autonomous
(n=53)
preoccupied
(n= 15)
1.7547 (. 20)
1.8034 (. 24)
1.7091 (. 32)
2,86
91
.
40
.
1349 (. 07)
1839 (. 10)
1418 (. 08)
2,83
2.47
09
.
5815 (. 18)
2,86
66
.
51
2,83
01
.
Contro II ing/negative
Note:
f-ratio
dismissing
(n=2 I)
Discipline/pLinishment
Positive niaternal
representation
df
.
.
5624 (23)
1277 (. 11)
two-tailed levels of significance
.
.
5230 (. 18)
1239 (. 12)
.
.
1201 (. 10)
.
.
98
207
Table 8.3 indicates that none of the factors remain significant at a level of p 05.
:! .
The Discipline/Punishment factor remains significant only to the extent of a trend (p =
09) This trend indicates that children of securely rated mothers remain those who use
themes of discipline and punishment the most while children of dismissing mothers
appear to use them the least. As there were no significant results, Scheff6's post hoc test
was not conducted.
A second post hoc analysis of variance was conducted to explore if it might be

distinguish
differences
between
for
to
the four-way system of classifying
possible
groups
mother's attachment.
This differs from the above test as it includes those adults
insecure-unresolved
be
Levene's
test was again observed.
to
the
of
category.
considered
However. no significant differences were found between the secure and insecure)
dismissing, insecure-preoccupied and insecure-unresolved classifications. The figures
for this analysis may be seen in Appendix D. 2 Caution is urged in interpreting even this
lack of results however as the group size for the unresolved group (n= 7) is too small to
make statistical analyses reliable.
8.3.3
Father data
The same questions that were addressedfor mother's data in the previous section
data.
Independent
father
t-tests
this
time
the
sample
were
repeated
of
asked
ai-c iio\\
deten-nine
father's
from
to
data
the
attachment
classification
ifthere
adult
-,
vere
the
usino
208
differences between the child narratives and father's attachment for any of the four
factors. Levene's test for the equality of variance was observed. The results are presented
in Table 8.4.
Table 8.4 : Means, SDS and levels of significance of four factors with father's AAl
(insecure
before
birth
vs secure)
of child subiect
Means (SD)
Factor
df
T-value
insecure
(n = 29)
secure
(n = 59)
1.7245 (. 30)
1.8023 (.22)
86
1.36
1936 (. 11)
1531 (.08)
83
1.82
86
00
.
06
.
5424 (.20)
.
1263 (. 12)
.
.
5427 (. 19)
.
1246 (. 11)
.
83
Positive maternal
representation
Note :p values are based on two-tailed levels of'signiticance
Significance
17
.
07
.
99
.
95
.
in
be
As
for
the above analysis.
see the table
can
There are no significant results
be
It
factor
to
for
significant.
the
(Table 8.4), there is a trend
insecure
judged
been
have
themes
fathers
of
more
use
who
that
children of
appears
This
fathers
judged
in
secure.
discipline and punishment their narrativesthan children of
instance,
In
data.
that
found
whose
the
children
that
mother
with
is the opposite result of
discipline
themes
and
punishment
of
more
used
secure
mother's v,-ereclassified
8.3.3a Controlling
indIcated
trend
there
and so a
a
was
non-significant
the
is
above
result
-Ithough
209
cautious approach was decided upon. As was done above with the data pertaining to
mother, a two-way analysis of variance was conducted to detennine if there is a
interaction
significant
effect of father's social class with father's adult attachment
for
factor. Table 8.5 below presents the means
the
classification
deviations
and standard
of the three groups of father's social class with father's
attachment measured during pregnancy.
Table 8.5 : Means and SD of father's social class and father's AAl attachment
classification with the Discii)line/Punishment factor.
Means (SD)
Discipline/
punishment
intermediate
professional and
managerial
occupations
insecure
(n= 4)
secure
(n= 18)
insecure
(n= 16)
secure
(n= 30)
insecure
(n= 7)
secure
(n= 8)
1553
.
(. 07)
1465
.
(. 08)
1799
.
(. 10)
1480
.
(. 08)
2468
.
(. 13)
1989
.
(. 09)
F-value
(df)
14
.
(2,82)
86
.
partly skilled
and skilled
As can be seen from Table 8.5, there is so significant interaction effect between
father's social class and father's adult attachmentclassification and, therefore, father's
father
following
for
in
be
the
the
attachment
analysesof
social classwill not controlled
data.
8.3.4 Post hoe analyses
Given that there Nvasa trend between one of the factors at the insecure versus
T-value
it
1.82.
(Discipline/Punishment;
07),
level
=
p=
categorisation
was
of
sccure
L.
210
decided to conduct a post hoc analysis to explore if it might be possible to distinguish
differences between the more specific categories of insecurity. In order to address this
question two analysesof variance were conducted. The results of the first are presented
in Tables 8.6 and 8.7. These tables present the possible differences between the
secure
insecure-dismissing
the
and
and the insecure-preoccupiedclassifications of father's adult
quality of attachment. Levene's test for the equality of variances was observed and
indicatedthat for the Controlling/Negativeand PositiveMaternal Representationfactors

a parametric one-way analysis of variance would be the appropriate test (Table 8.7).
However, for the Quality/Open Response and Discipline/Punishment factors there was
indication
(Levene's 3.55, p= 03 and 3.21, p= 04 respectively) that the
a significant
.
.
homogeneity of variance was skewed and so the equivalent non-parametric test, the
Kruskal-Wallis one-anova, was used (see Table 8.6).
Mean rank of the first two story stem factors grouped by three-way
interview.
father's
adult attachment
classification of
Table 8.6:
Mean rank
Factor
dismissing
autonomous
preoccupied
35.75
46.29
49.23
47.91
_Discipline/punishment
40.18
- F5
(0.7
3
Note :p values are based on two-tailed levels of significance
df
h-value
2.77
24
.
2.46
29
.
211
Table 8.7 : Means and SDS of story stem factors grouped by th-ree-wayclassification of
father's adult attachment interview.
Means (SD)
Factor
dismissing
(n= 18)
Positive maternal
representation
Note :p
.
.
5703
(. 21)
1591 (. 13)
autonomous
(n=59)
.
.
df
f-ratio
f-prob
preoccupied
(n=l 1)
5427 (. 19)
4969
.
(. 19)
2,85
47
.
1246 (. 11)
0785 (. 08)
.
2,82
1.52
62
.
based
levels
two-tailed
are
on
values
of significance
The abovetwo tablesshow that there are no significant differencesbetweenany

father's
factors
the
three
the
attachment classification.
and
groups of
of
A second post hoc analysis was conducted, as was done with the mother data, to
four
in
differences
distinguish
the
be
the
if
to
way
using
groups
possible
might
it
explore
differences
However,
father's
there
were no significant
attachment.
system of classifying
insecureinsecure-preoccupied
insecure-dismissing,
found between the secure and
and
2
D.
in
Appendix
be
The
table
seen
can
of
results
classifications.
unresolved
8.4 DISCUSSION
data,
final
the
regarding
This chapter presentedand examined
attachment
Interview
by
Adult
Attachment
the
as
assessed
classification
attachment
adult
parents'
discusses
This
birth
the
the
target
the
before
the
results
of
child.
section
of
conducted
found
between
differences
be
to
see
adult
if
could
aimed
analysis
parental attachment
"Ll
attachment classifications with regard to the responsesof the child's play narratives.
Severalpredictionshad beenmadein relation to the four factorsderived from

the
factor analysis. It was suggested that the semi
had
task
story
stem
parallels
-structured
with the Adult Attachment Interview in that both ask the subject to consider how they
would respond to questions or prompts that trigger the attachment or attachment-related
systems. In also considering the theories of communication in relation to attachment
theory it was suggested that those children who parents were considered secureautonomous,,that is,,could openly discussthe negative as well as positive featuresof their
childhood experience in a coherent manner, would be able to produce narratives that
were also coherent,,open and responsive. That is, it was predicted that the Quality/Open
Response factor would show differences in the adult attachment classifications.
However, this was not confirmed when looking at either mother's or father's attachment
Why
logical
line
classification.
might such a seemingly
of theory not prove to be the case
'7
One reasonmay be that, asmentionedin the introduction,mediatingfactorssuch

is
dyadic
dynamic
infant-parent
influence
the
the
and
as
relationship,which a
one, may
line from the parents' ability to tell their own history to their child's ability to construct
Oppenheim
(1995)
et all
make the point that construction of narratives is a
narratives.
interactive process between parent and child. Indeed, evidence that the infant-parent
factor
in
be
In
Chapter
the
a
mediating
is
provided
previous
chapter.
relationship may
Seven it xvasshown that the Quality/Open Response factor is significantly related to the
both
infant-father.
the
relationship
of
infant-mother
attachment
and
quality of
21)
Another reasonthat no significant results were obtained on this scale could be that
Interview
is
Attachment
Adult
the
the
task
and
story stem
comparing
an oN,er
in
thinking.
simplification
The AAI measures complex and subtle issues in discourse
MacArthur
Narrative
Coding
Scheme
does
The
the
and
narrative
and
not.
child"s
language and actions are not assessedin the same way at all.
It might prove an
future
discovering
to
the
to
study
assess
child
narratives
with a view
interesting
discrepancies between what a child might say and what they might enact for instance.
One prediction that was confirmed, with respect to the mother attachment only,
did
differences
factor.
This
discipline
to
the
scale
reveal
and punishment
was in regard
discipline
in
themes
tenns
of
and punishment
of
in the mothers' attachment classification
before
birth
Children
the
used
whose mother's were secure-autonomous
in the stories.
discipline
their
narratives.
themes
in
punishment
and
of
more
It may be that the ideas
internalisation
discipline
of parenting values are
and
regarding parenting styles of
investigation
Further
drawn
be
however,
assessing
this
yet.
cannot
conclusion
accurate,
is
beyond
information
be
this
disciplining
and
necessary
would
and
parenting
of
styles
large
to
the
Caution
number
the resources of the study at this time.
is also urged with regard
in
be
kept
type
mind.
the
one errors must
possibility of
of tests conducted and so
from
different
the ones presented
The next chapter of this thesis takes a
approach
developing
to
theory
the
profiles
to
of
attachment
apply
attempts
and
in previous chapters
for the story stem task.
214
CHAPTER NINE
INTERACTION
OF GENDER AND ATTACHMENT
215
9.1 INTRODUCTION
Chapter Four examined the relationship between the children's responsesto the
demographic
factors
including gender. Several main effects xere
narratives and several
found, not for the four factors, but for someindividual elementsof the narratl,,,es; such
variables as negative atypical response, physical aggression, physical punishment,
disciplining
father,
representations of
anger and adaptiveness of response showed
differences in the means of girls and boys.
Chapter Six explored the relationship between the child's attachment assessedin
infancy and the children's narratives and also found several important main effects. This
hypothesis
that the attachment classification of the child might
the
chapter will explore
have an additional effect beyond that of gender and will consider the combined
interaction effects of gender and attachment.
As mentionedin previouschapters,children's narrativesdemonstratethat by the

basic,
have
boys
acquired a
common
and girls across social classes
early school years,
differences
in
When
internal
for
describing
speaking about gender
states.
vocabulary
interactive
behaviours
fact
is
that
the
to
of
pre-school
in
there
support
evidence
ctieral,
,
be
be
tend
that
to
boys
girls
tend to more aggressive,active and impulsive and
children,
for
(Turner,
the
to
approval
need
adult
more passive, compliant and prone anxiety and
differences
in
to
Research
1991).
adults regarding their perspectives on
referring gender
differi
from
has
these
that
the
perspecti,,,,
ing
es
stem
child's
earliest
suggested
relationships
t,-
by
Because
the
are
oirls
mothered
a
person
of
same
gender
relationship.
of
experience
216
they come to experience themselves as less differentiated as boys and suggestthat a girls'
search for identity is through connection with others. The belief is that girls come to value
empathy, caretaking and orientation to the needsof others. In the caseof boys, the belief
is that becauseboys are treated as other from the start they must learn to represstheir
attachment to mother and that their 'relational potential' is inhibited by this process
(Chodorow, 1978; Tarullo, 1994).
In a study examining naturalistically gathered stories and narratives from

boys
found
Tarullo
(1994)
that the boy who acts or talks
and girls,
preschool
do,
his
tells
typical
stories about characters who
aggressively, or who
is not
of
gender.
She suggests that the presence of these 'outliers' among boys suggest that there are
different developmental trajectories for boys and girls from different social worlds.
However, she does not go into further explanations of what this different 'social' world
between
have
from
differing
If
the
genders
originated
perspectives
might consist of.
developmental
different
trajectories evolve
child's earliest experiencesof relationship and
for boys and girls from different social worlds, then one might reasonablely argue that
be
influential
the
quality of attachment.
early relationship would
one aspect of
However, this question, whether the quality of child-parent attachment may lead
hitherto
in
has
been
for
boys
different
the
question
a
neglected
to
and girls,
predictions
literature
(Turner, 1991).
One study that does point to gender and attachment
differences that appear in the preschool years in,.,,estigated concurrent links between
four-year-old
The
study
examined
children whose
attachment and peer interactions.
Strange
Situation
described
the
using
modified
in
attachment status Nvasmeasured
217
Chapter 6 (Cassidy and Marvin, 1989/92).
Turner (1991) found that individual
differences in observed behaviour in preschool were related to the quality of the motherfor
boys.
Insecurely
as
concurrently
assessed,
particularly
child attachment relationship
frequent
influence
boys
less
to
they
attempts control and
made
others,
showed
attached
helping
behaviour
and sharing and they showed more aggressive
such as
positive
behaviour. They also threatened, criticised and excluded others during peer games.
Insecurely attached girls showed significantly more positive expressive behaviour such
behaviour
This
than
secure girls.
was
as smiling and expressing pleasure with peers
be
seen as
interpreted as a way of avoiding conflict and/or exclusion and could
behaviour.
submissive
The aim of the current study is to investigate whether the differences found in
previous chapters regarding attachment classification and children's narratives might
both
Consideration
havean addedeffect when combinedwith gender.
attachmentand
of
following
in
in
the
this
arearesults
gender in terrns of the small amount of empirical work
predictions;
have
be
themes
boys
insecure
more
of
and
that
controlling
more
will
have
insecure
that
more
representations
will
girls
and
of
others
exclusion
aggressionand
factor.
Representation
Maternal
Positive
on the
218
9.2 METHOD
9.2.1 Sample and Design

The participants discussed in this chapter are the 89 children who have been
referred to throughout this text and are a part of the London Parent-Child Project sample.
Infant-mother attachment
When the sample of child subjects is examined in more detail with regard to
it
be
26.8
%
that
of the sample are girls
seen
gender and attachment classification can
20%
(n=26);
to
their
of the sample
mothers at one year of age
who were securely attached
boys,
insecurely
(n=20);
the
their
to
mothers at one year
of
attached
are girls who were
29.9 % were securely attached to mother at one year (n=29); and 22.7% of the boys were
insecurely attached to mother at one year (n=22).
Injbnt-father attachment
Looking closely at the sample in relation to fathers and attachment classification
itis found that 3 1.1% of the girls were securely attached to father (n=28); 16.7% of the
boys,
40%
(n=
15);
father
the
insecurely
to
of
were
at eighteen months
attached
girls were
insecurely
12.2%
(n=36);
fathers
and
were
their
to
months
eighteen
at
attached
securely
I).
I
(n=
attached
219
9.2.2 Procedure
The procedure for the administration and coding of the child narratives has been
described in detail in a previous chapter (Chapter Three) of this thesis.
9.3
RESULTS
The results are reported in two main sections. The first addressesthe results of
the analysesof infant-mother attachment classification and its possible interaction effects
five
the
the
outcome of
children's narrative responsesat
with gender on
years of age. The
infant-father attachment classification and possible interaction effects with gender are
four
factors
in
Within
the
to
the
eachsectionresultspertaining
secondsection.
addressed
derived from the story stemsare presentedas well as results for eachof the individual
variables.
9.3.1 Possible interaction effects of gender and infant-mother
attachment
A two-way analysis of variance was conducted to consider the interaction effects
for
12
the
the
outcome
child's genderupon
each
of security with mother at months and
Table
9.1
below
from
derived
factors
four
the
children's story stem responses.
of the
for
interaction
levels
the
two-way
F-values
of
significance
of gender and
the
and
gives
twelve
at
months.
mother
with
classification
attachment
220
Table 9.1: Means and SDS of the two-way interaction of gender and insecure -..,
's secure
for
four
factors.
the
attachment classification with mother at one year
Means (SDS)
insecure
Positive maternal
representation
secure
girls
(n= 17)
boys
(n= 2 1)
girls
(n= 23)
boys
(n= 27)
1.7187 (. 25)
1.6864 (. 28)
1.8520 (. 15)
1.8077 (. 27)
01 (1,87)
.
91
.
1269(. 06)
1394(. 07)
1700(. 08)
2016(. 12)
.
21 (1,84)
.
64
.
91
.
Negative/control ling
F-value (df)
5184(. 19)
.
5793 (. 23)
.
5053 (. 19)
.
5573 (. 16)
.
01 (1,87)
.
1705(. 13)
0864(. 08)
.
1285(. 12)
1224(. 12)
2.25 (1,84)
.
II
.
I
.
I
--
Table 9.1 above reveals that there are no significant results for any of the four
factors derived from the children's responsesto the story stems.
9.3.2
Post hoc analysis
As mentioned in the introduction, the question of whether the quality of childhas

been
boys
in
differences
lead
and
girls
about
to
predictions
parent attachment may
because
is
main
effects
Because
and
this
area
neglected
largely
a
such
neglected one.
a
be
hoc
further,
4),
Chapter
(see
analysis will
post
a
for particular variables were revealed
individual
interaction
the
of
investigate
variables
of
the
effects
possible
to
conducted
for
below
9.2
two-way
the
Table
the
analysis
of
results
shows
narratives and gender.
for
and
security
attachment
variables
representation
parental
and
variance ot content
for
by
followed
be
the
variables.
performance
This
results
will
gei-ider.
133
L---Jl
221
Table 9.2 : Means and SDS of two-way interaction of gender and insecure vs secure
for
12
months
attachment classification with mother at
representation variables.
Variables
Means (SDS)
insecure
boys
20)
girls
(n= 16)
Affection
Dishonesty
.
.
F-value (df)
1154 (. 13)
2.86 (1,84)
09
.
secure
(n=
boys
(n= 26)
girls
(n= 23)
1705 (. 16)
0682 (. 10)
.
1080 (. 15)
0955 (. 13)
.
0949 (. 09)
.
1189 (. 12)
43 (1,84)
.
.51
1146 (. 14)
Guilt/reparation
0455 (. 06)
.
0318
.
(. 08)
0711 (. 11)
.
0315 (. 07)
.
44 (1,84)
.
50
.
Negative atypical
response
0327
.
(. 07)
0868 (. 11)
.
0261 (. 06)
.
0815 (. 14)
.
001 (1,84)
.
97
.
Excludes other
0739 (. 11)
.
1136 (. 16)
1146 (. 11)
1958 (. 17)
40
.
(1,84)
52
.
Physical aggression
0341 (. 09)
.
2000 (. 19)
.
0830 (. 11)
.
1643 (. 13)
1.89
(1,84)
17
Positive atypical
response
0663
.
(. 14)
0469 (. 08)
.
0543 (. 07)
.
0827 (. 09)
.
1.15
(1,84)
28
.
Physical punishment
0341
.
(. 08)
1000 (. 09)
0672 (. 12)
.
1364 (. 17)
003 (1,84)
.
95
.
Disciplining father
0682 (. 07)
.
1273 (. 12)
0830 (. 07)
.
1434 (. 13)
001 (1,84)
.
97
.
Disciplining mother
2443
.
(. 12)
2136 (. 13)
.
3320 (. 15)
.
3042 (. 19)
.
002 (1,84)
.
96
.
1705 (. 14)
1045 (. 10)
1423 (. 13)
1294 (. 13)
87
.
(1,84)
35
.
1294 (. 16)
05 (1,84)
.
81
.
3007 (. 20)
.
48 (1,84)
.
48
.
Positive mother
Shame other
0966 (. 10)
.
Verbal punishment
2443
.
.
.
0818
.
(. 11) 1 2000
.
(. II)
(. 14)
.
.
1304 (13)
2925
.
(. 17)
.
.
is
however,
there
indicates
that there are no significant results,
Table 9.2 above
for
the variable of
trend
appearing
a
'affection'.
The narratives of insecure girls
instances
than
16;
secure girls
affection
of
(mean=. 17, sd=.
n=16) contain more
find
boys
looking
When
the
the
the
14,
narratives
opposite;
at
we
(mean=. 11, sd=.
n=23).
13,
11,
than
(mean=.
boys
of
affection
contain
more
instances
n=26)
sd=.
of secure
is
is
It
10,
06,
to
that
there
also interesting note
almost
n=20).
sd=.
insecure boys (mean=.
boys
-16)
(mean=.
11,
13
11:
(mean=.
in
the
n=2
sd=.
-,
and
girls
the
of
means
difference
no
222
be
is
judged
It
14;
insecure
to
the
secure.
sd=.
only
n=23) who are
children who are
for
differences
in
the
this variable.
showing
means
A two-way analysis of variance ws conducted for the perfonnance variables as
however,
there are no significant results for any of the performance variables.
well,
Figures for these calculations can be seen in Appendix E. 1.
Because of the large number of tests conducted concerning the individual variables,
the possibility of type one errors increases and must be kept in mind.
9.3.3 Possible interaction effects of gender and infant-father attachment
A second series of two-way analyses of variance was conducted to consider the

father
interaction
and the child's gender upon the
effects of security with
possible
below
factors.
9.3
F-values
four
for
Table
the
the
shows
and
story stem
each of
outcome
levels of significance for the two-way interaction of gender and attachment classification
father
at eighteen months.
with
insecure
interaction
SDS
Means
9.3:
the
two-way
Table
vs secure
of genderand
of
and
four
factors.
for
father
18
the
at months
attachmentclassification with
F-value (df)
Means (SDS)
insecure
Disc Ipline/punishment
Negative, 'control ling
E
os tIve maternal
tpe riesentation
secure
girls
(n= 14)
boys
(n= 11)
girls
(n= 24)
boys
(n= 34)
1.6952 (. 27)
1.6791 (. 34)
1.8516 (. 14)
1.7711 (. 27)
27 (1,82)
.
60
.
1795(. 09)
1485(. 11)
1338 (. 07)
1892(. 10)
3.13 (1,79)
08
.
5082 (. 22)
.
6134(. 28)
.
5043(. 18)
.
5583 (. 16)
.
28 (1,82)
.
59
.
1643 (. 13)
1091 (. 08)
1420(. 13)
0992(. 13)
.
43 (1,79)
.
51
.
.
L---
223
Table 9.3 indicates that there are no significant results, however. unlike the
results of the analysis with infant-mother data, there is an indication of a trend for the
Discipline/Punishment factor. The narratives of secure boys (mean = 19 sd = 10,
.
,
.
insecure
(mean=.
18, sd =. 09, n=13) contain more themesof discipline
n=33) and
girls
Insecure
boys
(mean = 14, sd =. 11, n= 10) and secure girls (mean =
and punishment.
.
13, sd 07, n= 24) show fewer punishment and discipline themes in their narratives.
.
9.3.4
Post hoc analysis
As was done above with the infant-mother data, an analyseswill be conducted to

investigate the possible interaction effects of individual variables of the narratives and
gender.
Table 9.4 below shows the results for the two-way analysis of variance of
for
attachment security and gender. This
content and parental representation variables
followed
by
be
the results of the analysis of the performance variables.
wil
224
Table 9.4 : Means and SDS of two-way interaction of gender and insecure N"ssecure
father
for
18
at
months
attachment classification with
reoresentation variables.
Means (SDS)
Variables
F-value (df)
insecure
Affection
Dishonesty
secure
boys
(n= 10)
girls
(n= 13)
boys
(n= 33)
girls
(n= 24)
1288 (. 14)
0992 (. 13)
.
09 (1,79)
.
75
.
1000 (. 10)
0758 (. 07)
.
1047 (. 13)
79 (1,79)
.
3)7
.
0455 (. 11)
.
0606 (. 10)
.
0275 (. 07)
.
43 (1,79)
.
51
.
(. 02)
0735 (. 13)
.
0354 (. 08)
.
0918 (. 13)
.
03 (1,79)
.
85
.
1538 (. 14)
0909 (. 08)
.
0682 (. 08)
.
1873 (. 19)
6.03 (1,79)
1608 (. 17)
1259 (. 17)
Guilt/reparation
0490 (. 07)
.
Negative atypical
response
0070
.
.
.
1091 (. 08)
01**
.
Excludes other
Physical aggression
0769
.
(. 10)
2455 (. 22)
.
0606 (. 11)
.
1625 (. 13)
88 (1,79)
.
34
.
Positive atypical
response
0839
.
(. 14)
0848 (. 11)
.
0508 (. 08)
.
0597 (. 09)
.
02 (1,79)
.
87
.
Physical punishment
0490
.
(. 08)
0909 (. 09)
.
0530 (. 12)
.
1405 (. 15)
47 (1,79)
.
49
.
Disciplining father
0699
.
(. 06)
1273 (. 13)
0795 (. 08)
.
1433 (. 13)
01 (1,79)
.
90
.
Disciplining mother
3077 (. 12)
.
(. 16)
2837 (. 17)
.
33 (1,79)
.
56
.
1553 (. 14)
1267 (. 12)
81 (1,79)
.
36
.
1047 (. 13)
4.95 (1,79)
02*
.
(. 19)
1.19 (1,79)
Positive mother
Verbal punishment
Note:
05
.
2545
.
(. 20)
2879
.
1678 (. 12)
0818 (. 09)
.
(. 13)
1091 (. 16)
0606
.
(. 07)
(. 17)
2538
.
(. 16)
2028
.
1-2937
Shame other
*p>
(. 12)
**p > 01
.
2182
.
.
.
.
.
2755
. .
The
9.4
(Table
interesting
content
above).
There are two rather
significant results
level
05.
both
'shame
of
p<
at
a
significant
'excludes
are
other'
and
other'
variablesof
.
both
that
of these variables are contained within the
Perhapsthis is not surprising given
9.3).
The
boys
Table
factor
(see
trend
that
secure
above
showed a
15
14,
I
3
3)
(mean
the
19,
the
sd
=
n=
girls
=.
use
insecure
18,
and
3)
n=
sd = .
(mean = .
.
.
-'
225
insecure
boys
boys
Insecure
(mean
than
theme excludes other' more
and secure girls.
07,
08,
(mean
24)
09,
10)
08,
=.
sd = .
n=
use this theme in
and secure girls
= .
sd = .
n=
their narratives less. The picture is different for the theme of 'shame other. In this case,
there is no difference between the mean for the boys. Both insecure (mean =: . 10, sd =. 16,
is,
13,
3
3)
boys
There
(mean
10,
'shame
10)
=.
sd =.
n=
other' equally.
use
and secure
n=
however, a marked difference in the insecure and securegirls use of this theme. Insecure
06,
(mean
13,
13)
20,
(mean
this
theme
than
sd
secure girls
=.
use
more
sd =.
n=
=.
girls
24).
07,
n=
=.
for
figures
the
below
9.5
the
two-way
Table
the
of
variance
analysis
of
presents
father
by
at
eighteen
with
classification
attachment
and
gender
variables
performance
months.
226
Table 9.5 : Means and SDS of two-way interaction of gender

and insecure vs secure
attachment classification with father at 18 months with the performance variables
(n=89).
Variables
Means (SDS)
insecure
F-valuc
(di)
secure
girls
(n= 14)
boys
(n= 11)
girls
(n= 24)
boys
(n= 34)
Role of parent
1.7857 (. 44)
1.5702 (. 65)
1.8561 (. 37)
1.8102 (. 42)
60(1,82)
.
A-)
Child's understanding
of conflict
1.0455 (. 14)
1.0413 (. 17)
1.0833 (. 11)
1.0241 (. 18)
50(1,82)
.
48
.
Directness of
performance style
1.8506 (. 27)
1.7521 (. 41)
1.9659 (. 09)
1.8877 (. 29)
02(1,82)
.
87
.
Responsivity to
examiner
2.5519
2.4876
2.7803
2.7727
06(1,82)
.
79
.
Involvement of
examiner
1.4091 (. 27)
1.4876 (. 30)
1.4167 (. 27)
1.4305 (. 29)
21 (1,82)
.
64
.
Investment in
2.0779
(. 83)
2.2562
2.2311
2.1845
(. 61)
43 (1,82)
.
51
.
2208
.
(. 15)
2149 (. 06)
.
3342 (. 19)
.
41 (1,82)
.
52
.
Adaptiveness of
response
1.6039 (. 38)
1.4463 (. 27)
2.0000
(. 40)
1.6471 (. 33)
1.27(1,82)
26
.
Narrative coherence
2.5065
2.7025
2.8939
(. 39)
2.6257
3.31 (1,82)
07
.
(. 60)
(. 64)
(. 92)
(. 30)
(. 62)
(. 39)
performance
Denial
Control
(. 57)
1472 (. 22)
(. 55)
2879 (. 16)
.
(. 57)
1515 (. 27)
0505 (. 12)
.
1141 (. 21)
36(1,82)
.
54
.
Joy
4307
.
(. 20)
5207 (. 27)
.
5303 (. 29)
.
5062 (. 24)
.
84(1,82)
.
)6
.3
Anger
0736
.
(. 05)
0441
.
(. 05)
0758 (. 07)
.
0463 (. 04)
.
00(1,82)
.
97
.
Distress
0281
.
(. 05)
3'1 (. 04)
0278 (. 04)
.
0187
.
(. 03)
48(1,82)
.
48
.
Concern
0671
.
(. 12)
0354
.
(. 12)
0455 (. 05)
.
2.66(1,82)
10
6515
.
(. 27)
6212
.
(. 27)
Anxiety
.03
0248
.
(. 02)
5455 (. 31)
.
5668 (. 29)
.
13 (1,82)
A trend for narrative coherence and concern is shown in Table 9.5 above. Girls
be
judged
father
been
have
to
secure
with
at eighteenmonthsgive narrativesthat are
\vho
2.89,
')9,
24)
judged
(mean
be
insecure
than
sd
(mean =
=
=.
n=
to
girls
coherent
more
71
.
*71
11
227
2.5 0, sd =. 57, n= 14). On the other hand, boys j udged to be insecure (mean = 2.70, sd
boys
(mean
2.63,
55,
11)
than
coherent
narratives
secure
more
slightly
=
sd
give
n=
-.
is
intriguing,
it
be
difference
34).
Although
57,,
this
the
that
result
must
noted
n=
-.
between the mean of these two groups is very low.
9.4 DISCUSSION
This chapter explored the possibility of an interaction effect when gender and
in
task.
the
to
stem
completion
story
relation
attachment classification are combined
Very little research has been conducted regarding the combined effects of gender and
attachmentand so this study makesa notable contribution.
interaction
is
effects
there
The first result to take note of that
were no significant
infant-father
infant-mother
factor
to
or
found for any of the four
scaleswith regard either
boys
between
differences
Four
and girls
Chapter
that quantitative
data. It was noted in
look
boys
girls
will
and
found
the
aggression,
of
that,
exception
with
and
were not often
in
be
This
1994).
to
the
(Tarullo,
case
appears
measures
alike on most socio-affective
differences
is
for
may
the
gender
A
results
non-significant
reason
this instance. possible
is
It
in
that
gender segregation a group
is argued
be i-clated to being
a social group.
inhibition
display
the
gender
stereotypical
of
or
encourage
may
that
groups
process ad
1991).
(Turner,
bellaviours
228
Although there were no significant results for the factor scales, there was one
trend noted for the Discipline/Punishment factor indicating that boys secure with father
at eighteen months used more themes of discipline and punishment in their narratives at
five.
This
be
age
could
suggestive of secureboys being more able than insecure boys to
internalise adaptive models of discipline. In support of this, Oppenheim (1997) found
that disciplining representationsin the narratives are most indicative of children, whether
boys or girls, to have the least likelihood of behaviour problems. When it
comes to the
in
this study, however, the opposite seemsto be true. That is, insecure girls used
girls
more themes of discipline and punishment in their narratives than secure girls. The
for
this are not so apparent. Could it be that insecure girls have
possible reasons
internalised parental values regarding discipline in relation to their fathers in a more
?
Of
negative and self-punitive way
course, this is speculation as there is no firm
for
disciplining
evidence,
example parenting and
styles, on which to rely.
Again, it must be mentioned that, becauseof the large number of tests conducted
increase
individual
the
type
this
possibility
of
one
the
errors
variables,
and
concerning
discussed
below.
kept
be
When
individual
the
to
the
results
must
in mind with regard
interesting
variables were examined some potentially
results emerged that will need
tIurther investigation in future studies. For example, in relation to attachment with mother
for
'affection'.
trend
twelve
was noted
at
months, a
Affection is scored when there is an
kisses,
hugs,
'Good
compliments,
or
praise,
such
as
girl". An interesting
instance of
pattern emerges; insecure girls use more affection themes than secure girls and secure
boys usemore affection themesthan insecureboys. One explanation for the higher use
be
Turner's
finding
by
insecure
to
may
related
that
girls
insecure
girls exhibit
of affectiori
229
more positive behaviour such as smiling and expressing pleasure as a way of avoiding
her
findings
for
Another
to
this
of
relates a significant result
conflict and/or exclusion.
insecure
boys
found
Turner
during
threatened,
that
study;
criticised and excluded others
in
it
is
However,
boys
the
our sample,
secure
peer games.
and insecure girls who refer
to this theme more often. Again, the results are not easy to explain. The last significant
for
individual
is
'shame
Girls
father
the
that
variables
of
other'.
result
insecurewith
at
18 months showed markedly more use of this theme than any of the three other groups.
In conclusion, the above results suggestthat there may be an interaction effect of

future
but
that
work is need to clarify this possibility.
gender and attachment
230
CHAPTER TEN
PROFILE ANALYSIS
2-31
10.1 INTRODUCTION
The study that will be described in this chapter was not included in the original
The
follows
in
thesis.
this
that
procedure
of
plan
arose the courseof study as a result of
discussions regarding the sample of children and the various ways in which they
individual
in
to
stories
relation to their attachmenthistory. The data contained
responded
within this chapter are presented purely in the interests of exploration and provoking
further thought regarding attachment organisation in preschool children.
The previous chapters have looked at the narrative task in relation to various
found
For
between
example, associations were
measuresof adaptation and attachment.
the Child Behaviour Checklist and the Control I ing/Negative factor and such individual
lack
physical
aggression
and
of anxiety
scales as positive maternal representation,
(Chapter 5).
With regard to the attachment data, differences were found between
father
both
factors
insecure
mother and
on several
with
children classified as secureand
and
individual
scales,
particularly
the
Quality/Open
Response
and
Discipline/Punishment factors (Chapter 7). In one instance, a post hoc analysis was
factors
discover
four
to
the
story
order
per
in
if
story stem
conducted calculating each of
differences
between
distinguish
be
to
secure and
able
any one particular stem might
7).
Chapter
(see
insecure children
The factors used in these analyses were calculated using the variables across all
different
dilemma
However,
the
story
stems
contains
of
a
each
or
conflict
stories.
eleven
in
Chapter
Three,
As
mentioned
was
using of all the scales
Ior the child to consider.
232
Coding
Scheme
in
factor
Narrative
Macarthur
the
the
included in
analysis was not done
it
as would have been misleading. Potentially interesting variables such as compliance
included
factor
the
and non-compliance were not
in
analysis as these could have different
meanings within the specific context of each story. For example, non-compliance in the
Bathroom Shelf suggests a prosocial motivation while in Mother's Headache it might
defiant
suggesta
or self-interested motive. Although compliance was reliably rated the
internal consistency across stories is low and this was another reason for not including
it in the factor analysis.
The author's experience in administering hundreds of the stems to the five-yearled

found
be
to
that
the
to
observations
some of
old children
children
particular stories
in
instance,
Lost
Keys
For
than
the
others.
story, which two parents
more anxiety raising
keys,
loss
The
Burnt
Hand
to
the
arouse
great
of
seemed
uneasiness.
are arguing over
deal
of concern in the children and yet some were able to respond
story also raised a great
in a coherent and organised manner while others were not.
The construct of quality of attachment has not yet been applied to the preschool
years with
as thorough
developmental
a
is
available
perspective as
for infancy
has
been
The
1992).
(Crittenden,
on understanding the meaning of the
emphasis
indeed,
has
from
infancy
this
thesis
the
to
years
and,
preschool
continuity of patterns
its
The
in
how
that
of
chapters.
several
question
remains
of
question
examined just
how
its
function
in
is
functions
the
the
and
preschool
years
integrated
into
attachment
dc\'elopment of the preschool child (Cicchetti, Cummings, Greenberg & Marvin, 1990).
It was decided that, along side of analvsing the data described in chapters Four to Nine,
23'3
interesting
be
to
and potentially
another way of approaching the story stems might prove
in
issue
in
the
the
construct of attachment the preschool years.
of
useful
addressing
Therefore, one of the aims of this chapter is to develop a theoretical profile for
how a child assessedin each of the four attachment classifications might respond to each
battery.
The
is
the
the
secondaim of
chapter to apply theseprofiles to the sample
story of
in
how
known
infant
to
the
the
order
assess
accurately
profiles match with
of children
classification.
10.2 METHOD
10.2.1 Sample and Design
The sample discussed in this chapter is the same sample of 89 five-year-old

in
detail.
The
in
Chapter
Three
described
have
been
children's
previously
children who
in
Phase
the
second
assessed
was
one
year
at
age
mother
with
classification
attachment
Waters
&
Wall,
Blehar,
(Ainsworth,
Situation
Strange
Ainsworth's
of the project using
longitudinal
Project
for
III
Phase
five
the
1978). At age
of
years, the children returned
The
Story
Stem
Battery.
MacArthur
the
children's story completions were
and completed
Narrati-ve
Coding
Scheme
MacArthur
Chapter
Three
described
the
using
in
coded as
(MNCS).
234
10.2.2 Procedure
Development of the profile
As stated, this study involved developing a profile in which each of the variables
Coding
Scheme
in
MacArthur
dilemma
Narrative
the
to
the
of
was considered relation
four
A
then
the
made about which of
contained within each story. prediction was
insecure-avoidant,
insecure-resistant
(ie.,
secure,
categories of attachment classification
high
low
five-year-old
(or
insecure-disorganised)
score
or
present
children
would
and
Spilled
Juice
For
for
the
the
example,
story
variables in each story.
each of
or absent) on
happen
the
tell
to
the
child character accidentally
when
will
and
what
show
child
asks
knocks juice onto the floor. The rater would then make a judgment regarding which type
be
It
high
coherence.
could
on say, narrative
of attachment classification would score
high
from
on narrative coherency while an
theory that a securechild would score
argued
low.
insecure-disorganisedchild would score
Three raters first met to discuss the criteria for making the judgements. The three
having
battery
familiar
the
to
the
stories
administered
story
stem
with
raters were all
from
the
ages of
children
man,,,
5-10 years.
The raters were also familiar with
developmental
infant
All
the
sequelae
attachment.
theory
of
raters
were
and
attachment
blind as to the attachment classification of the children.
235
The raters discussedtheories mentioned in the opening chaptersof this thesis such
base
behaviour,
conununication and affect regulation theories of thinking about
as secure
differences in attachment and how these might be applied to the children and story stem
task. Prototypical profiles for each attachment pattern were created. These are described
below.
10.2.3 Prototypical profiles
Secure childprofile
direct
style
of
presentation
and
a
communication
of
open
evidence
storyline
coherent
both
task
to
and
examiner
- responsive
in
high
the
denial
the
little
anxiety
stories
particularly
story,
within
conflict
of
or
no
Reunion.
Separation
Hand,
Burnt
Keys,
Lost
and
of
distress,
joy,
instances
ie.,
there
concern
and
anger,
of
are
expressed
affect
of
range
-a
anxiety
behaviours
lack
or
statement
of
controlling
dolls
utilised
parent
-a
balance of positive and negative representations of parents
disciplining
parents
of
representation
appropriate
empathy,
affection,
themes
as
such
- positive
injury
and negative atypical
lack of neoative themes such as physical aggression,
response.
236
Avoidant childprofile
direct
style of presentation
- evidence of a coherent storyline and a
to
to
task
the
opposed
responsive
as
examiner
denial
in
Lost
Keys,
Burnt
the
of
of
conflicts
evidence
within
some
stories,
particularly
Hand, Separation and Reunion stories
little
lack
of
affect
expressed,
range
particularly
a
of negative affect expressed
lack
behaviour
of
controlling
statements
or
-
dolls
little
of
parents
use
lack
parental
representations
and
a
of
negative
parental
representations
positive
disciplining
of
representation
parents
lack
themes
of
negative
themes
self
and
excludes
as
compliance
such
-
Resistant childprofile
incoherent
stories
and
coherent
of
mix
indirect
of
performance
style
the
task
to
highly
to
opposed
as
examiner
responsive
in
lack
to
the
of
a
resolution
resulting
stories
within
preoccupation with conflicts
stories
-a
distress,
as
anger and
range of affects expressed, particularly negative ones such
anxiety
behaviours
statements
and
controlling
- some
disciplining
of
parents
lack
representations
of
-
237
Disorganised childprofile
direct
indirect
incoherent
a
stories and mix of
and
styles of presentation
-
lack
to
of
resolution
stories
to
task
responsive
and examiner
behaviour
of
evidence
controlling
statements
and
-
denial
of conflicts within stories
- evidenceof
distress
fear
negative
affect
expressed,
particularly
and
injury
themes
negative
such
as
physical
aggression,
and negative atypical responses
-
To give another example, the rater would consider that in the Burned Hand story
the dilemma is of the child receiving an injury as a result of disobedience.The rater then
infant-mother
herself
high
him
the
type
question, which
of
attachment will score
or
asks
low,
for
?
The
this
type
story
of attachment
rater
on negative atypical responseand which
then marks on the scoring sheet that insecure-disorganised children will score 'present'
tIor this variable and that securely attached children will score 'absent' (see Appendix
F. 1). This procedure is followed for all variables.
10.3 RESULTS
10.3.1 Inter-rater
reliability
238
Inter-rater reliability was calculated on all variables in each section of the
MacArthur Narrative Coding Scheme. That is, all the variables in the content, parental
representation and performance and affect sections of the manual were used in
developingthe profiles. Each rater consideredeachvariable in relation to the dilemma

in
presented the story and then made a judgement as to whether a subject might score
high (or present) or low (or absent) according to their attachment classification (see
Appendix F. 1).
There are 18 content variables that proved to be reliably rateable (from Chapter
3) and II stories, therefore, there are a total of 198 judgements to be made. For the
judgements
be
for
66
to
the performance
there
are
made,
parental representation section,
for
decided
decisions
It
judgements
198
99
the
to
affect variables.
was
and
variables
into
division
the three periods of presentation, narrative
the
the
stories
of
original
retain
for
this analysis.
transition
phases
and
Table 10.1 below presents the variables in the stories in which all three raters
both
decided
three
It
the
those
that
all
raters
agreed
where
on
variables
only
agreed. was
for
further
be
in
If
any
analyses.
story
would
used
each
one
present and absent rating
looks at the first line of the table below, for example, all three ratersjudged that the
Mother's
Headache
be
Bathroom
Shelf
the
present
in
and
'affection'
would
vartable
low)
in
for
(or
(B)
those
absent
tor
and
stories
avoidant
children
children
secure
storics
(A).
239
Table 10.1 : High (or present)and low (or absent)ratings per variable per story where all 3 raters agr ed.
Story Stems
Si
MH
3C
BH
LK
SS
SN
RN
BS
EX
Affection
Compliance
h
I
Dishonesty
h
I
Guilt/reparation
h
I
B
A
Non-compliance
C
B
Negative atypical
response
h
I
Excludes other
h
I
Physical aggression
h
I
Denial
h
I
Responsivity
h
I
Role of parent
h
I
Negative paternal
I
representation
Negative maternal
I
representation
h
I
Control in presentation
phase
h
I
D
B
D
B
Control in narrative phase
h
I
D
B
D
B
Control in transition
phase
h
I
D
B
D
B
Joy in presentation phase
B
A
B
A
B
A
Jo% in narrative phase
B
A
B
A
B
A
Note:
II
A- insecure-avoidant
C- insecure-resistant
B
A
B
A
B
A
B
C
A
B
D
B
BT
D
B
D
B
D
B
D
B
A
B
D
B
D
B
A
B
A
B
D
B
B
A
A
B
A
B
A
B
A
B
B
A
B
A
B
A
B
A
D
B
D
B
D
B
D
B
D
B
D
B
D
B
D
B
B
A
B
A
B
A
B
A
B
A
B- secure
D- insecure-disorganised
B
A
240
What Table 10.1 also illustrates is the main problem that resulted and that is a
lack of agreement between raters.
As mentionedabove,it was decidedto count only thoseinstanceswhereall three

both
the present and absent rating of each variable for each story. Table
raters agreed on
10.2 below presents the percentage of agreement for each section.
Table 10.2 : Percentage of agreement for content, parental representation and

performance variables.
percentage of agreement
9.09%
Content variables
18/198
Parental representation variables
8/66
12.12%
Performance variables
14/99
14.14%
Affect variables
13/198
6.56%
As can be seenfrom the figures above (Table 10.2) agreementis very low, so low
between
As
is
be
the
three
it
fact
there
raters.
a
that must
really no agreement
said
in
beyond
between
this
lack
the
should
not
proceed
study
the
raters
of agreement
result of
point.
However. as was stated in the introduction, this is an initial investigation based

inquiry
if
line
future
this
to
of
worth
pursuing
as
a
see
is
assumptions,
theoretical
on
decided
it
to
the
the
to
and
proceed
exploration
As
continue
next
step.
was
such,
StUdY.
241
10.3.2 Profile analysis
A single profile or template for each attachment classification was constructed

as
described above in section 10.2.2 using the variables where all three
raters agreed. That
is, a template for how secure, insecure-avoidant insecure-resistant
insecureand
,
disorganised children are predicted to respond to each of the stories, was constructed.
The four completed templates were then each compared to the four attachment
groups of
in
the sample. Pearson correlations were conducted and each
children
correlation was transformed
into a z-score,
Fisher's
formula,
using
so that the
be
correlations could
added together. A meansof the Fisher's z scoreswas taken for each
in
these
attachment group and
are presented Table 10.3 below.
Table 10.3 : Mean correlationsof z-scoresbetweenthe profiles and the four attachment
2rounsof children.
Secure
template
Secure children
Avoidant children
Resistant children
Disorganised children
L.
.
.
.
.
1303
1227
1166
1536
Avoidant
template
.
.
.
.
Resistant
template
Disorganised
template
1542
0355
.
1406
0392
.
0899
.
1266
0380
.
1186
1464
0358
.
0586
.
1007
As can be seen from Table 10.3. the mean correlations of the z-scores are very
failed
to correlate with the attachment group for which
templates
the
that
low sIgnit'viii(g,
242
they were designed. Not only that, the templates were just as likely to correlate with anv
other attachment group. For instance, the avoidant template was just as likely to correlate
disorganised
with secure,,resistant and
children's responses.
It had been the intention to analyse the difference in the means of the groups,
however,asthe correlationsare so low this was not worth pursuing and ftirther analysis
be
would
meaningless.
10.4 DISCUSSION
The aim of the chapter was to develop a theoretical profile for how a child
battery.
in
infancy
The
the
to
story stem
with mother might respond
second
assessed
in
these
to
the
to
then
the
profiles
sample of children
apply
order to
chapter was
aim of
how
accurately the profiles matched.
assess
The most obvious result of the experiment, and the first issue that must be
between
There
failure
is
to
the
raters.
are two main ways
achievereliability
addressed,
first
is
based
faulty
failure
that
the
this
the
the
profiles are
study,
of
on
of considering
itself
in
five
how
the
organisation
might manifest
stories of
theory about
attachment
year
have
been
that
the
the
second,
prototypical
profiles
not
clearly
old children, and
I
operationalised.
%Vhileit is acknowledged that much work Iis needed to understand
beyond
this
this
thesis,
infancy,
there
indeed
one
is
goal
of
children
in
is
attachment
243
in
Chapters
(reviewed
I
empirical evidence
and 2 and within this thesis) that certain
elements of the narratives, such as coherency, responsivity and disciplining parental
representations,are relatedto attachmentthe favouredexplanationfor the failure to gain

is
this
time
that the profiles have not been clearly operational i sed.
reliability at
For example, regarding content variables, the three raters tended to agree that
secure children would show the least amount of negative atypical responsesand that
insecure-disorganisedchildren would show the highest number of these bizarre

felt
be
for
This
dealing
to
true
responses.
was
particularly
stories
with attachmentrelated
issues such as injury in the Bumed hand story or the Separationand Reunion stories. The
insecure-disorganised
have
internalised
that
children would
assw-nptionwas
no
organised
in
face
issues
injury
fall
back
the
to
of
raising
such
anxiety
as
and
strategy on which
bizarre
responses.
separation and so give
In looking at the performance variables section, the raters tended to agree that
by
looking
the
the
to
smiling,
at
examiner
responsivity
more
securechildren would show
It
task.
that
face
to
the
also
agreed
was
completion
story
and responding
examiner's
by
least
their
the
averting
gaze
of
responsivity
amount
show
avoidant children would
to
the
task
from
the
affect
and
responding
neutral
primarily
showing
examiner,
away
reluctantly.
However, operational i sation of this code in the N4NCS manual may not be
instance,
For
the
this
to
the
to
chapter's
goal.
responsivity
of
purposes
appropriate
both
the
the
to
the
of
child
responding
examiner
and
elements
contains
examiner scale
244
task. From an attachment theory point of view it may be more accurateto separatethese
it
be
instance,
For
theoretically
could
concepts.
arguedthat securechildren would be able
to use their internalised model of secure attachment to enable them to respond to both
the task and the examiner in an open and easy manner. However, children with insecurefocusing
by
their attention on the
of
attachment might respond well
avoidant patterns
task rather than responding either positively or negatively with the examiner which would
is
This
them
to
the
engagewith
experimenter.
similar to what insecure-avoidant
require
Situation,
been
doing
Strange
have
the
avoiding contact with any
observed
in
children
figure
the
exploring
environment.
on
and
concentrating
attachment
or
substitute
real
However, these ways of thinking about the ways in which the different patterns
development
in
did
of profiles which
not result successftil
of attachment might respond
in
infancy
with mother.
correlated well with any of the groups identified
is
described
in
MNCS
the
individual
manner
Perhapsassessing
variables of the
does
by
A
give
a
sense
not
analysis
variable
variable
approach.
not a particularly useful
in
dilemma
that
the
the
the
to
story
resolving
of the overall strategy or approach
further
for
One
this
type
do.
on
work
to
suggestion
attempting
are
profiles
prototypical
the
be
consideration
with
profiles
the
of
to
operationalisation
clarify
of analysis would
discussion.
the
above
given to
does
The
N4NSC
is
the
not make
affect section.
Another area worth expanding
themselves
that
the
between
or
coding
affect
subject
child
in
emotion
codincy
distinction
a
tN
From
the
of
experience
administering
the
is
personal
story.
within
portraying
the subject
245
narratives , it is quite a different experience to observe a child who is distressed while
is
fully
the
task
to
and
observing
one who
attempting complete
engaged in the task and
having the character in the story portraying distress.
In a recent paper yet to be published (von Klitzing et al, 1998), useful comments
and changesto the narrative coherence scale have been made but have yet to be applied
to a sample of children. The changes they describe may also be useful in separating the
is,
that the scale assess
that
that
the
are confusing,
narrative coherencescale
elements of
both coherenceand addressingthe conflict.
There is a further problem with using the variable by variable approach. It does
injury
themes
that
the
as
and physical
such
so-called negative
problem
not address
In
in
be
this
the
age.
other
play of children
used quite appropriately
aggression might
be
is
dealt
how
into
theme
it
that
more
with
could
taken
a
consideration
not
words, is
Hand
in
Burned
is.
For
the
story
might
theme
the
child
a
than
example,
what
important
for
injury.
However,
in
hand
bums
their
a score
resulting
mention that the sibling also
hospital
both
the
to
the
had
and
the
injury
children
the
rush
parents
sample
one child in
for
In
dealt
also
scored
present
a
child
another,
empathetically.
and
swiftly
with
was
her
hand
burning
doll
the
however,
without
repeatedly
that
child
enacted
subject
injury,
incident
her
to
the
was provided.
aid and no resolution
the parents coming to
in
ideas
described
is
believed
this chapter regarding
the
that
In conclusion, it
it
however,
be
first
has
that
the
must
emphasised
merit.
profiles
prototypical
constructitl(-)
I
be
that
to
tackled
important
of
Is
reliability.
problem
and most
The following chapter
246
investigations
data
dealt
in
describe
final
the
the previous
the
of
with
will
statistical
chapters of this thesis.
247
CHAPTER ELEVEN
MULTIVARIATE
REGRESSION
ANALYSIS
.248
11.1
INTRODUCTION
This chapter describes the final statistical investigations of the data described in the
previous chapters of this thesis.
Firstly, a discriminant function analysis was conducted in order to ascertain if the

be
belong
basis
the
to
the
of
attachment
membership
subjects
could
predicted on
category
Secondly,
hierarchical
their
a
performance on particular variables.
cluster analysis was
of
homogeneous
based
to
try
groups of cases
and identify relatively
on
conducted in order
in
1997).
Ideally
(SPSS,
the
characteristics
selected
would result
selected characteristics
identifying four separate clusters, each composed of one of the four types of attachment
be
between
the
If
to
analysis could
expected
casesof insecurity,
security. not able identify
insecure
identify
two
cases.
to
clusters of secure and
11.2
METHOD
11.2.1 Subjects
London
Parent-Child
discussed
the
this
part
of
chapter
are
a
The participants
in
been
has
this
text.
to
throughout
referred
Project sample. which
249
11.3 RESULTS
The results are divided into two sections according to the type of statistical test
first
The
conducted.
section describes the results of the discriminant function analyses and
the second presents the results of the hierarchical cluster analysis.
11.3.1 Discriminant
function analysis
In order to determine if a statistical distinction between insecurely and securely

by
Strange
(as
Situation) exists which can be accounted for
the
attachedchildren
measured
differences
in
by
the means of the children's responses on several factors,
or measured
discriminant function analyses were carried out. In order to determine if there was any
for
single variable or smaller group of variables with predictive powers
attachment
function
discriminant
the
analyses was used.
stepwise method of
classification
Stepwise discriminant function analyseswere carried out on both mother and father
data using two different sets of discriminating variables. The first set used was the four
tactors (Quality/OpenResponse,Discipl Ine/Pulnishment,Controlling/NegativeandPositive

Maternal Representation) derived from the factor analysis in Chapter Three as the
discriminating variables to predict grouping by infant attachment to mother and then to
4-thther and hy the parental adult attachment classification. The second set of variables used
individual
discriminating
the
the
factors
stories as
across
variables to predict
the
agnoregated
250
grouping by infant attachment status and parental adult attachment classification.
The critical minimum F to enter the step-wise analysis is 3.84 and the maximum
F to remove from the analysis is 2.71.
11.3.1.1
Predicting groups by infant security classification with mother
For infant security with mother it was determined that 58-82% of the children
identified
insecure
in
infancy
five
or secure
at
with mother were correctly
classified as either
below).
in
(see
Table
11.1
For
by
their
themes
the
security with
narratives
used
yearsof age
373,
Lambda
[Canonical
R
Wilkes
8609,
=.
=.
mother, one canonical variable was extracted
df = 2, X2 = 12.28, p=. 002]. This one canonical variable indicated that two of the factors
(Quality/Open Response and Discipline/Punishment) contributed highly to the canonical
infant
best
the
security.
predictors of
variable and were
When the variable for father's social class was also added to the discriminating
factors
G.
The
1)
(see
Appendix
difference
little
two
same
resulted
variables, very
.
best
infant
Discipline/Punishment)
Response
the
(Quality/Open
were
predictors of
and
R=
[Canonical
security
381, Wilks' Lambda =. 8543, X, = 12.59, df = 2, p=

.
%
57.65
the
cases grouped correctl
of
),
predictin, (:
001],
.
251
Table 11.1: Discriminant function analysis grouped by infant attach-ment security with
mother.
Actual Group
I No. Of Cases
I Predicted group membership

11
Group 1
1 36
141.7%
15
1 21
58.3%
Group 2
49
14
28.6%
35
71.4%
Ungrouped cases
101
0%
.
100.0%
Percent of "grouped" casescorrectly classified: 58.82%
From the table above (Table II- I) it can be seen that the analysis placed 35 of the
49 secure children in the secure group. Fourteen of the 49 secure children were placed
insecure
looking
15
36
When
insecure
in
the
the
predictions,
only
at
of
group.
incorrectly
incorrectly
insecure
21
in
the
children were
placed
correct group while
children were placed
in the secure group.
This table indicates that the four factors were better at placing the
insecure,
in
in
than
the
the
correct group
who were
children secure infancy with mother
inore otten identified as secure.
252
11.3.1.2
Predicting groups by infant security with mother using factors per storN
For infant security with mother using the factors calculated per story, it was
detirmined that 64.77% of the children classified as either insecureor securein infancy
identified
five
in
by
their
the
themes
at
used
years of age
with mother were correctly
below).
(see
Table
11.2
For
security with mother, one canonical variable was
narratives
352,
8760,
[Canonical
R
Wilkes
Lambda
=.
=.
extracted
X2
10.85, df = 2, p= 004]. The

.
is
factor
Quality/Open
Response
important
to
the
the
canonicalvariable
contribution
most
in two of the stories (Spilled Juice and Biscuit Tin).
When the variable for father's social class was also added to the discriminating
father's
did
G.
The
1).
(see
Appendix
difference
variable,
social class,
resulted
variables, no
factor,
into
Therefore,
for
the
the
same
analysis.
entry
not meet the statistical criteria
Quality/OpenResponse,in the sametwo stories(Spilled Juiceand Biscuit Tin) was the best
infant
security.
predictor of
253
Table 11.2 : Discriminant function analysis using the factors calculated per story as
discriminating variables grouping by infant attachment security with mother per story.
Actual Group
I No. Of Cases

12
Group 1
38
17
44.7%
21
55.3%
Group 2
50
10
20.0%
40
80.0%
From the table above (Table 11.2) it can be seen that the analysis placed 40 of the
50 securechildren in the securegroup (80.0%). Ten of the 50 securechildren (20.0%) were
insecure
looking
insecure
in
When
17
incorrectly
the
the
at
predictions,
only
group.
placed
insecure
21
(55.3%)
in
38
the
children
correct group while
were
of
incorrectly placed in the securegroup. As for the results above using the four factors over
better
factors
indicates
the
that
this
table
calculated per story were
at placing
all the stories,
insecure,
in
in
infancy
than
the
the
group
the children secure
correct
who were
with mother
more otten identified as secure.
11.3.1.3
Predicting groups by infant security classification with father
When using only the four factor scales to predict group membership with regard to
factors
father
level
the
the
months,
none
eighteen
of
at
met
minimum
infant security with
for
As
it
the
the
no
variables
qualified
analysis
to
analysis.
was abandoned.
ot'tolerance enter
254
However, when the variable for father's social class was entered in addition to the
tIour factor scales, it could be determined that 71.59% of the children classified as either
insecure or secure in infancy with father were correctly identified at five years of age (see
Table 11.3 below). In this case it is not the themes used in the children's narratives that
best
For
father,
father's
is
the
predictor.
security with
is
it seems
social class the best
infant
[Canonical
R= 23 1, Wilks' Lambda = 9465, X2 = 4.15, df = 1,
security
predictor of
.
.
04].
.
Table 11.3 Discriminant

father.
Actual Group
function analysis grouped by infant attachment security with

No. Of Cases
Predicted group membership

12
Group 1
25
Group 2
63
0
0%
.
0
0%
.
25
100.0%
63
100.0%
Table 11.1 above shows that the analysis placed all 63 of the secure children in the
25
insecure
A
that
the
all
of
children were
securegroup. positive result until one considers
father
in
infancy
indicates
The
father')
in
table
that
the
group.
securewith
s social
also placed
father
infancy
identifies
their
the
secure
with
in
as
regardless
children
of
responses
all
class
be
This
result may the result of a type one error, particularly considering
to the story stems.
the next section.
255
11.3.1.4
Predicting groups by infant security with father using factors per story
The picture is quite different when attempting to predict group membership by

infant security with father when the individual stories are entered into the analysis.
When
father's social class is not taken into account, it was detennined that 78.75% of the children
in
father
identified
five
insecure
infancy
or secure
with
were correctly
at
classified as either
below).
in
(see
Table
11.4
For
by
their
the
themes
security with
narratives
used
years of age
father by story, one canonical variable was extracted [Canonical R= 438, Wilks' Lambda
.
is
df
0003].
The
16.46,
2,
XI
8075,
the
of
canonical
variable
comprised
=
=
p=
= .
.
loaded
factor
Positive
Matemal
Response
loaded
Quality/Open
the
and
negatively
positively
Representationfactor in the same story, the Exclusion Story, and best predicts infant security
father.
with
When father's social class is added to the equation, the percentage of the children
factors
in
The
G.
1).
(see
Appendix
81.25%
two
the
increases
same
to
correctly classified
7782,
Lambda
Wilks'
R
470
[Canonical
best
=
=.
predictors
one story remain the
.
,
18.80, df = 2, p= 000 1].
.
X2
256
Table 1IA: Discriminant function analysis using the factors calculated per story as
discriminating variables grouped by infant attachment security with father.
Actual Group
I No. Of Cases

12
Group 1
23
9
39.1%
14
60.9%
Group 2
57
3
5.3%
54
94.7%
57 (94.7%) securechildren in the securegroup. Only 3 of the 57 securechildren were placed
incorrectly in the insecure group. When looking at the insecure predictions, 9 of 23 (39.1%)
incorrectly
insecure
14
in
the
placed
children were
correct group while
better
factors
indicates
This
that
the
stories
are
table
calculated across
in the securegroup.
insecure
infancy
in
father
than
group
group membership
at correctly predicting securewith
membership.
11.3.2
Predicting groups by parental adult attachment classification

determine
function
to
discriminant
As stated above,
analyses were also conducted
insecure
between
distinction
those
Ia
classified
parents
were
children
whose
it statistical
be
for
Interview)
Attachment
by
Adult
the
(as
could
accounted
or
measured
secure
N-ersus
factors.
in
the
differences
by
the
children's
on
several
responses
of
means
measured
257
11.3.2.1
Predicting groups by maternal AAI classification
When looking at mother's adult attachment classification (as measuredby the AAI),
67.42%
identified
five
detirmined
that
the
of
children
were correctly
at
years of age
it was
by the themes used in their narratives (Table 11.5 below). Again, just one canonical variable
[Canonical R=
237, Wilks' Lambda = 9434, X2 = 4.85, df = 1, p= 02] was extracted

.
.
best
factor
Response)
(Quality/Open
that
the
was
predictor of mother's
comprisedof one
adult security classification.
However, the picture changeswhen the variable for father's social class is added (see
Appendix G. 1). The Quality/Open Responsefactor is no longer the best predictor; the factor
for Discipline/Punishment becomesthe best predictor [Canonical R ==.245, Wilks' Lambda
The
df
02].
1,
5.06!,
9397,
XI
percentageof the grouped cases correctly classified
= p =.
=
=.
drops to 59.30%
258
Table 11.5: Discriminant function analysis grouped by mother's AAI.
Actual Group
I No. Of Cases

11
Group 1
1136
13
36.1%
163.9%
23
Group 2
1 53
6
11.3%
188.7%
47
47
of
be
the
that
placed
11.5)
analysis
(Table
seen
it can
From the table above
53
placed
Six
were
children
the
secure
in
of
the
group.
secure
53
the
secure children
13
of
only
looking
the
When
predictions,
insecure
at
in
the
group.
insecure
incorrectly
23
were
children
in
insecure
while
the
group
correct
36
placed
the
children were
factors
four
the
were
that
This
table
indicates
in
the
group.
secure
incorrectly placed
the
than
the
group
correct
in
secure
are
mothers
better at placing the children whose
secure.
as
identified
often
more
insecure who were
11.3.2.2
Predicting groups by maternal AAI classification using factors per story
When looking at mother's security of attachment using the factors calculated per
69.77%
determined
it
the
that
of
children whosemother's were classifiedas either
story was
birth
before
five
(see
Table
their
correctly
were
identified
at
years
of
age
insecure or secure
11.6 below). For mother's security, again only one canonical variable was extracted
[Canonical R =. 417, Wilks' Lambda = 8256,
X2
15.80, df = 3. p= 00 1]. There are three

.
259
factors
(Quality/Open Response, Discipline/Punishment and Positive Maternal
separate
Representation) in two different stories (Biscuit Tin and Mother's Headache) that best
AAl
mother's
classification.
predict
When looking at mother's adult security of attachment using the factors calculated
into
father's
figures
taking
the
as
well
as
account
story
social
per
class,
remain the same (see
Appendix G. 1).
discriminating variables grouped by mother's AAl.
Actual Group
I No. Of Cases

12
Group 1
35
15
42.9%
20
57.1%
Group 2
51
6
11.8%
45
88.2%
51 securechildren in the securegroup (88.2%). Six of the 51 securechildren (11.8%) were

looking
insecure
When
in
insecure
15
the
the
at
predictions,
group.
only
placed incorrectly
insecure
20
in
(57.1%)
35
the
group
while
children
correct
were
of'
ilICOITeCtIN'
for
four
As
factors
the
the
the
results above using
secure group.
over
placed in
factors
indicates
better
that
the
table
this
calculated
per
story
were
at placing
all the stories,
the
secure
in
correct group than the insecure,who were
are
mother's
the children whose
260
identified
more often
as secure.
11.3.2.3
Predicting groups by paternal AAI classification
It was not possible to predict group membership for father's attachment classification
father's
including
the
social class, qualified and the so analysis was
variables,
as none of
abandoned.
11.3.2.4
Predicting groups by paternal AAI classification using factors per story
When looking at father's security of attachment using the factors calculated per story
it was determined that 68.24% of the children whose father's were classified as either
insecureor securebefore their birth were correctly identified at five years of age (see Table
11.7 below). The one canonical variable extracted [Canonical R =. 234, Wilks' Lambda =
9450, X' = 4.66, df = 1, p =. 03], is comprised of the Discipline/Punishment factor in the
Biscuit Tin story and best predicts father's AAI classification.
Taking into account father's social class makes no difference (see Appendix G. 1).
261
discriminating variables grouped by father's AAI.
Actual Group
I No. Of Cases
11
Group 1
1 27
2
7.4%
192.6%
25
Group 2
1 58
2
3.4%
196.6%
56
Table 11.7 shows that the analysis placed 56 of the 58 secure children in the secure
incorrectly
(3.4%)
58
the
Only
(96.6%).
the
two
in
placed
were
children
secure
of
group
insecuregroup. When looking at the insecure predictions, only 2 of 27 children were placed
in
incorrectly
(92.6%)
insecure
the
25
secure
placed
were
children
the
while
group
correct
in
better
factors
indicates
the
that
at
placing
Again,
story
were
per
table
calculated
this
group.
insecure,
in
the
father's
than
the
who were more
correct group
the children whose
are secure
identified
as secure.
often
11.3.3 Hierarchical
cluster analysis
in
identify
to
try
hierarchical
and
relatively
A
cluster analysis was conducted order
(SPSS,
1997).
Ideally
based
the
liornogeneous groups of cases
on selected characteristics
four
identi,
in
separate
clusters,
each
composed
Ing
of
result
Nvould
characteristics
selected
identify
between
If
to
four
able
cases
not
of
security.
types
attachment
of
the
one of
II
262
insecure
be
two
to
the
clusters
of
secure
and
expected identify
insecurity,
analysis could
cases.
A first dendrogram was generated using the four factors (Quality/Open Response,
Discipline/Punishment, Control ling/Negative and Positive Maternal Representation) as the
dendrogram
in
detennine
if
The
to
the clusters
analysed
order
was
selectedcharacteristics.
by
(see
G.
2).
One
looking
Appendix
can see
of groups generated were meaningful or not
identified.
further
As
G.
2
Appendix
there
that
a result any
were no meaningful groups
at
be
meaningless.
analyseswould
A second dendrogram was generated using the two factor scales that proved
in
Chapter
7
(Quality/Open
from
insecure
in
distinguishing
groups
secure
significant
Response and Discipline/Punishment).
Would a regression analysis be able to identify
dendrogram
factors
?
This
from
homogeneous
two
these
second
was
groups
relatively
further
identified
however,
there were again no meaningful groups
and
analyses
analysed,
were abandoned.
11.4
DISCUSSION
This chapter described the final statistical investigations of the data described in the
thesis.
this
of
previous chapters
263
First conducted was the discriminant function analysis to ascertain if the category of
be
belonged
basis
to
the
the
subjects
could
predicted on
of their
attachment membership
discriminant
function
The
that
the
results suggest
performance on particular variables.
is
however,
interesting
to
the
tending
identify
majority of children as secure,
some
analysis
difficulties
One
be
because
in
the
there
of
could
were many more children
result emerged.
the sample secure with mother and father than insecure.
One thing that has emergedand has been consistent throughout the previous chapters
is that
the Quality/Open Response and Discipline/Punishment factors emerge as the
discriminating variables. This seemsto be evidence to the validity of these two factors being
important in distinguishing secure from insecure groups.
The cluster analysis failed to generateany meaningful clusters and so further analyses
were abandoned.
Further discussion of these results follow in the summary and discussion chapter.
264
CHAPTER12
CONCLUSIONS
265
12.1
INTRODUCTION
Early in this dissertation (Chapter 2) research in the field of child development

investigations
have
to
pursued empirical
into the child's
and psychoanalysis was shown
despite
developmental
into
infant
much
research
and adult
mental representations and,
be
done
there
to
much
is
work yet
regardingunderstandingthe
patternsof attachment,
inner
has
This
the
thesis
world of preschool children.
attempted to add
organisation of
to the information regarding the nature and continuity of attachment patterns and how
they might manifest in preschool children.
This final chapter chapterboth surnmarisesand brings together the findings of this
indeed
into
the
that
there
the
are
internal world of
young child and shows
investigation
reflections of infant patterns of attachment in preschool children's narratives.
The
findings,
the
the
confirmed and unconfirmed predicted
chapter provides an overview of
for
further
investigations
hoc
research.
and also makes suggestions
post
Nearly one thousand story completions offered in response to emotionally

focus
in
dilemmas
the
the
story stems were
of this study and many
presented
challenging
disconfirm
in
to
tests were conducted order confirm or
predictions made. As a large
number of tests were performed, the chance of type one errors occurring increased so
this was kept in mind and care was taken in reporting significant results.
266
12.2
CONFIRMED
FINDINGS
12.2.1 Overview of reliability and data reduction
As stated,this thesis has attemptedto add to information with regard to assessing

the representationalworld of the preschool child, particularly with respect to attachment
theory, and has confirmed many of its predictions.
One primary aim was to explore the potential of a new research tool, the
MacArthur Story Stem Battery, as a method for eliciting/gaining accessto the child's
had
Previous
that
the
research
world.
shown that
internalworld andassessing quality of
children as young as three years of age could provide coherent resolutions to the
dilemmascontainedwithin the stemsand that parentsinfluencedthe construction of the

investigation
however,
thorough
of the psychometric properties regarding
a
narratives,
the tool had beenabsentin theseprevious studies.
Thus, an important contribution of the study has been the psychometric analysis
Stem
Story
Battery
MacArthur
MacArthur
the
the
associated
coding
scheme,
and its
of
Narrative Coding System. It was demonstratedthat, not only could the content themes
but
MSSB
that the performance
the
produce a set of reliable and consistentscales,
of
In
and
consistent
scales.
reliable
addition, those scalescould
codescould also produce
basis
factor
be
be
the
and,
on
stories
of
a
across
analysis,
then aggregated
aggregated to
I
form meaningful constructs with high internal consistency.
267
Four meaningful factors were constructedthat suggest four relatively independent
factor
largely
data
identified
by
dimensions
This
the
analysis.
was
a
psychological
were
driven process which yielded those aggregate scales and yet they reflect conclusions of
factors
found
both
Distinct
theory
the
and research.
were
which characterised
past
first
factor
The
the
the
of
children's
narratives.
performance
is comprised
contentof and
of narrativecoherency,responsivityto the examiner,child's understandingof the conflict
directness
of performance style and relates to the narrative and communication
and
history
the
that
of the child's emotional communication and the
emphasises
viewpoint
(Oppenheim
bases
&
Waters,
narrative
construction
underlying children's
cognitive
1995). The discipline and punishment factor suggests a view of parents as authoritative
being
harshly
This
disciplinary
seemsto correspondsto research
punitive.
without
and
contrasting authoritative parenting with authorl
describes
II
parenting,
which
firm
high
demanding
rules and
standards,
mature
authoritative parents as imposing
behaviourwithout resort to excessivephysical punishment. The controlling and negative
factorpicksup elementsof both the content and performanceof the children's narratives
that offers a way to addressboth the child's representations of thematic content as well
features
behavioural
that
style
are suggestiveof
of clinical
as addressingan aspectof
discipline
&
Cicchetti,
The
Clyman
1992).
Toth,
(Buchsbaum,
and punishment
samples
tactor seemsto focus on aspectswhere child and parent are in conflict. Considerationof
is
factor,
fourth
that
the
indicative of maternal warmth and
a relationship variable
formulations
by
the
to
of
internallsationof values making a
affection, is related most
in
to
the
to
order
parental
values
please
parent and maintain a
accept
willing
more
child
(Grusec,
1997).
pleasurable relationship
268
12.2.2 Child Adaptation - CBCL
The aim to replicate studies comparing the results of the children's story stem
Child
Behaviour
to
the
shed
a
measure
of
child
adaptation,
well-establi
responses
Checklist, proved partially successful. It was predicted that, as this scale contained
difficulties
more closely related to clinical
elementsassociatedwith children experiencing
brought
it
be
to
that
are
child
guidance
clinics,
associatedwith the
problems
would
Controlling/Negative factor.
ratings of the child.
This hypothesis was confirmed when looking at father's
As expected, children who scored higher on the dimension
behaviour,
physical aggression and negative atypical responses,
suggestiveof controlling
tended to be rated by their father as having externalising behaviour problems. This
correlation predictedwas significant.
12.2.3 Attachment
Bowlby's
importance
in
the
of child-parent communication
ideas about
development
are revealedin the recent trend of assessments
of attachment,
attachment
described
here.
Underlying
the
much of this work
including useof children'snarratives
key
internal
idea
that
characteristicsof children's
working models influence
is the
interpreted
have
described
investigators
in
Chapter
the
results
of
studies
and
narratives
Two as supporting the notion that young children construct internal working models with
&
Waters,
1995).
This
(Oppenheim
literature
to
the
study
adds
to
and
respect attachment
by
MacArthur
Story
Stem
the
Battery
that
theory
new
evidence
providing
this
supports
269
can provide a window on earlier attachment relationships.
There were two approachesto the investigation of the internal world with respect
to attachment.The first looked at the ability of attachmentclassifications,both child and

factor
to
the
the
scales.
predict
scores on each of
adult,
The second approach, a
multivariate regression analysis,explored the question of which previously assessed

independently
influenced
five
attachmentvariables
children's narrativesat
years.
Although there were differencesin the results between mothers and fathers, it was
the sametwo factors that gave positive results confirming the validity of these two factor
for
this sample. With respect to the possible influence of the mother-child
scales
father-child
it
to
the
relationship, was anticipated that there
relationship as opposed
be
This
the
greater
of
mother-child
relationship.
a
possibly
influence
assumption
would
be
to
confirmed.
appears
While the possibility of type one errors occurring must always be kept in mind,
factor
from
factor
derived
four
the
the
two of
scales
analysiswere repeatedlyshown to
be significantwith regardto earlierattachmentclassifications. Thesetwo factors, quality
discipline
broadly
both
how
the
and
and
punishment,
of
response
andopenness
represent
the childrencommunicatetheir storiesand one particular aspectof what the children talk
about. In this sampleof middle-classchildren , whether or not a child was securely
by
level
the
predicted
to
was
of children's scores at age five
attached mother at one year
from
dimensions
the
and
emotional
openness
of
coherence
the
of
story
and
examining
on
differences in their representations of discipline and punishment. This was explained in
270
describes
from
line
theory
terms of a communicationperspectiveof attachment
a
which
in
easy and open emotional communication
infancy to the ability to construct open and
issues
internalisation
to
the
and
of parental
coherentnarratives about affectively charged
line
developmental
linking
in
The
sensitive and responsive maternal caregiving
values.
the first year of life to children's representations of a disciplining parent at age five may
be indicative of how the meaning of parental sensitivity and responsivity evolves as a
function of the child's developmental age and stage.
The discriminant function analysis revealed that these same two factor scales,
important
discipline
and punishment, are
in
quality and openness of response and
60%
mother,
as
much
as
of the time, providing
security
in
infancy
with
predicting group
further evidenceto support the theory that these two dimensions are important elements
in
the
organisation of attachment preschool children.
in
Further evidence in support of claims regarding the importance of the early

for
father
in
terms of promoting
relationship with parents suggestsan additional role
This
and
responsive
of
organising
narrative.
suggestion
open
ways
emotionally
coherent,
father
finding
based
the
that
the
month
assessment
eighteen
of
attachment
with
on
also
is
dividing
the
quality
and
openness
of
response
with
scale
when
producedan association
Children
had
children.
the sampleinto insecureand secure
who
securerelationshipswith
direct,
father
to
tended
produce
coherent and open narratives. By
their
in infancy
father-child
the
insecure
early
in
relationship,
predicted
attachment
narratives
an
contrast,
indirect,
less
less
responsiveand
emotionally open. The story
that were more incoherent,
271
findings
discussed
below.
this
specific
and are
support
12.2.4 Story specific findings
Another contribution of this study as been the exploration for knowledge about
individual
stories. Different storiespull for different themesand the post hoc
aspectsof
few
best
distinguish
that
to
conducted
indicated
a
particular stories are
analysis
able
insecure from secure children in this sample. On the dimension of quality and openness
of response, Spilled Juice, Burned Hand, Exclusion, and Biscuit Tin are the stories that
discipline
dimension,
In
Spilled
Juice
the
terms
of
and
punishment
significant.
were
and
Mother's Headache were significant. Both Spilled Juice and Mother's Headache place
the child in potentially conflictual situationswith mother and so It is not surprisingthat

they relate to the discipline and punishment theme.
The discriminant function analysis was able to predict group membership with
by
larger
the
analysis was
story providing evidence that
a
percentage of successwhen
for
different
do
be
themes
that
and
particular
pull
stories
stories
may
individual
more
discriminant
function
The
than
to
others.
analysisalso
related attachmentorganisation
in
it
is
Spilled
Juice
the
the
that
openness
quality
and
of
response
indicated
and Biscuit
Tin stories that best predict attachment group membership.
When looking at the child's attachment to father in relation to
stories
for
different
best
distinguish
than
those
form
that
stories
mother
secure
it appears
272
Shelf
Exclusion
Bathroom
Reunion
The
and
stories proved
insecure children.
story,
insecure
in
The
significant at predicting
exclusion story,
and secure classification..
particular,
proved highly significant and the discriminant function analysis also
finding
by
being
this
able to predict attachment group membership 79% of the
supported
time.
This exclusion story is of theoretical interest. Psychoanalysis has placed great

Oedipal
the
the
importance of
conflict and Its resolution. The children In this
emphasison
five
height
Oedipal
the
the
and considered at
of
phase. The results
sample are age
indicate
Story
father
Exclusion
that
the
to
children
secure
relationships
with
regarding
five
to
this
to
that show
produce responses
story at age
at eighteen months are able
joyous,
the
perform their stories in a
conflict, are more coherent and
understanding of
direct manner and who are more responsive to the examiner. Is there something about
in
infancy
better
father
that
the
to
to
enables
child
organise their
a secure attachment
thoughts and feelingsregardingan important developmentaland psychologicalconflict

four
later
?
Oedipal
the
years
conflict
suchas
It hasbeensuggestedthat narrative assessments

of children's internal worlds are
how
childrenconstruct those narrativesabout affective themesand
asmucha measureof
they
to
those
others
as
are assessmentsof internal representations and
communicate
&
Waters,
This
has
(Oppenheim
1995).
study
provided evidence to
working models
be
line
that
to
think
thinking
may
the
useful
it
about narrative assessments,
of
support
Stem
Battery,
Story
MacArthur
as measuring children's abilities to construct
such as the
27)
laden,
narratives about emotionally
personal topics and to share those narratives with
between
links
based
The
theory
a child's experience of sensitive, open
are
others.
crucial
feelings
of security, well-modulated affect and coherently
communication with parents,
figures.
organisedinternal working modelsof self and attachment
Post hoc analysesproved interesting. The scale describing quality and openness
found
be
to
significant when the sample is viewed from the original
of responsewas also
three-way classification of secure, insecure-avoidant and insecure-resistant. It appears
that the children with a history of security to mother and/or father give responsesthat are
highest on quality and openness and that the children with a resistant history give
it
least
be
However,
kept
the
quality
and
openness.
must
responseswith
in mind that,
becausethe group sizes of the insecure-resistant group was so small, these results must
be
It
necessary to apply these scales to populations
would
remain questionable.
before
disorganised
be
children
reliable conclusionscould
containingmore resistantand
drawn.
The findings for father's social classmerit further investigationas so few of the
from
lower
However,
the results
socio-economic
come
classes.
actually
current sample
is
better
that
analysis
at predicting which attachment group a child
a story specific
suggest
80%
(over
belong
time).
the
to
of
will
274
12.3
UNCONFIRMED
FINDINGS
12.3.1 Demographic characteristics
Other than the one main effect of father's social class on the discipline and
factor,
between
demographic
factors
there
no
associations
were
punishment
such as
language
found.
This
temperament
the
abilities
age,
and
child's
and
child's narratives
in
function
that
the
the
this
story
completions
of
children
sampleare not a
suggests
of
their age, languageabilities or temperament. With respect to temperament, it may be that
however,
be
One
is
to
the
task;
this
cannot
ruled out.
suggestion
narrative
it not related
for further study would be to assesstemperament concurrently with the narratives.
12.3.2 Regarding the Controlling/Negative
and Positive Maternal Representation
factors
Despite the psychometric analysis in Chapter Three successfully yielding four

I
tactor scales, only two of those scales proved to provide consistent positive results as
describedabove.One of the remainingtwo factors, describedas the controlling/negative

tactor, contains negativeelementsof both how the child performed and content themes
It
the
to wluch
child referred. was predictedthat this scalewould identify those children
infancy,
however,
did
be
ho
this
to
in
the case.
not
prove
Nxv, xvereinsecure-disorganised
Despite one association with externalising behaviour problems on the ChIld Behaviour
275
Checklist, the factor failed to distinguish secure from insecure children. This was
hypothesised
It
terms
the
the
explainedin
of
non-clinical nature of
population studied.
is
that this factor would be relatedto children showing more clinical disturbancesand this
by
Child
Behaviour
the
to
the
association
externalising
supported
problems on
is
Checklist. Suggestionsfor fiirther research would be to apply thee dimensions to clinical
populations.
Indeed, attempts to do have yielded supporting evidence (Buchsbaum,
Toth, Clyman, Cicchetti, & Emde, (1992).
The fourth factor, Positive Maternal Affection, contains two positive elements of
the content of the children's narrative. This factor failed to distinguish secure from
This
or
adult
attachment
assessment.
measure
of
child
was
insecure children on any
for
When
lack
themes.
terms
the
stories that
positive
pull
of stoneswhich
of
explained.in
have pulled for positive themes are included, positive themes tend to result more often.
For example, two stories which have been used in subsequent studies, Looking for
Barney and Gift Giving, have both elicited scorings of affection high above the mean for
their sample (von Klitzing, Kelsay & Emde, 1998).
One element of this factor, positive maternal representation, was found to be
behaviour
Child
Checklist.
the
to
scale
of
associated mother's rating on internalising
However,the direction of the associationwas not expected.Children who scoredhigher
likely
have
behaviour
to
internalising
on positive maternal representations were more
difficulties. This could possibly be a type two error, particularly in view of the fact that
be
(Oppenheim,
Emde
&
Warren,
have
1997).
to
true
the
opposite
shown
other studies
from
the associationsof this variable to the attachmentdata
Also, therewas no evidence
276
to support anything other than a spurious result at this time.
Clearly,further researchis neededto better explore the possibilitiesthis technique

for
different
have
Work
larger
populations.
work
with
populations
containing
n-ight
with
insecure
battery
to
the
the
of
children
might
yield results as
subgroups
usefulness
might
have in differentiating the insecure groups from one another. Also, work with different
types of maltreatedsamplesmight have implications for use of the battery as a diagnostic
tool.
12.3.3 Modified Strange Situation
The five year assessmentof attachment produced no associations to any of the

four factor scaleseither with mother or father. This was discussedin the chapter and the
lack of results suggest that the Cassidy and Marvin (1992) scoring of the modified
Strange Situation lacks validity.
12.3.4 Gender
This study has also answereda call for information about sex differences(von
Klitzing. Kelsay & Emde, 1998). There were no main effects found for gender alone on
hoc
Post
factor
four
for
analysis
some
revealed
predicted
scales.
the
associations,
of
anNF
277
example, in aggression.
There were also no associations regarding the interaction of gender and
factors
father.
four
Individual
the
with either mother or
variables
attachment and
but
further
the
that
conclusion reached was
study was needed.
revealed some results
Mary Ainsworth herselfhas expresseddisappointmentthat attachment researchers

have gone on to do research with the Strange Situation rather than looking at what
happensin the home. One suggestion for future research concerning attachment patterns
do
just
home
is
'field
(Ainsworth
the
that,
to
to
and
return
work'
in preschoolchildren
from
"ons
be
Observati
Marvin,
1995).
the
then
above
resulting
III
might
also
applied
and
to improve the operationalisationof the profiles of the various attachment classifications.
CONCLUSIONS
12.4
The above results support Bowlby's hypothesis that an individual's attachment

for
development
have
both
their
that
subsequent
consequences
and
prenatal
relationships
have
that
quantifiable
of
attachment
reflections
of
attachment at
assessments
and intant
I
278
five years of age. These findings were achieved on the basis of administering the welldeveloped
Story
Stem
MacArthur
Child
Checklist
Behaviour
the
and
newly
established
Battery.
A limitation of the study was the homogeneous sample population and so the
limited.
larger
Again
the
studies containing
results are
subgroups of
general1sability of
high-risk
and
also
samplesare necessary.
insecure classifications
It is perhaps the capacity of the MacArthur Story Stem Battery and the
identify
Coding
System,
how
MacArthur
Narrative
to
the
accompanyingcoding scheme,
identify
to
the content of the
to
coherency, as well as
children respond with respect
internalised
discipline
to
elements
of
which underlies the
responses with respect
be
findings.
The
takes
the
that
study
suggestion
it may
present
useful to
up
significant
think of narrative assessmentsas measuringchildren's abilities to construct narratives

(Oppenheim
laden,
&
topics,
to
then
share
and
others
personal
with
about emotionally
Waters, 1995).
279
REFERENCES
Achenbach,T.M. (1983). Manual for the Child Behaviour Checklist and Revised Child
Behaviour Profile. Burlington: University of Vermont, Department of
Psychiatry.
Ainsworth, M. D. S., Blehar, M. C., Waters, E. & Wall, S. (1978). Patterns of attachment:
A psychological study of the strange situation. Hillsdale, NJ: Erlbaum.
Ainsworth, M. D. S. & Marvin, R. S. (1995). On the shaping of attachment theory and

S.
Mary
D.
Ainsworth.
An
research:
interview with
In E. Waters, B. Vaughn, G.
Posada and K. Kondo-lkemura (Eds.), New growing points of attachment

Monograph of the Society for Research in Child Development, 60 (2-3, Serial
No. 244).
Bates, J., Freeland, C. & Lounsbury, M.
(1979). Measurement of infant difficulties.
Child Development, 50,794-802.
Bishop, & Edmundson (1987). The Bus Story
Bowlhv, J. (1969/82). Attachment and loss: Volume 1. Attachment. New York: Basic
Books.
280
Bowlby, J. (1973). Attachment and loss: Volume 2. Separation. New York: Basic Books.
Bowlby, J. (1980). Attachment and loss: Volume 3. Loss, sadnessand depression. New
York: Basic Books.
Blatt, S.J. (1995). Representational structures in psychopathology. In D. Cicchetti and

S.L. Toth (Eds.), Rochestersymposium on developmentalpsychopathology: Volume
6 Emotion, cognition and representation. University of Rochester Press.
Bretherton,,1. (1987). New perspectives on attachment relations: Security, communication

internal
In
Osofsky
J.
(Ed.
),
Handbook
development,
infant
and
working models.
of
New York: John Wiley & Sons.
Bretherton, 1. (1995a). Attachment theory and developmental psychopathology. In

D. Cicchetti
S.
(Eds.
Rochester
developmental
L.
Toth
),
and
symposium on
Volume
6:
Emotion,
University
cognition and representation.
psychopathology:
of
Rochester Press.
Bretherton, 1.(1995b). The origins of attachment theory: John Bowlby and Mary Ainsworth.
In S.Goldberg, R.Muir and J.Kerr (Eds.), Attachment theory: Social, developmental,
London:
The
Analytic
Press.
and clinical perspectives,
Bretlierton, 1. (1995c). A communication perspective on attachment relationships and

internal working
models.
In E. Waters, B. Vaughn, G. Posada and K. Kondo-
281
Ikemura (Eds. ), New growing points of attachment. Monograph of the Societyfor
Research in Child Development, 60 (2-3, Serial No. 244).
Bretherton, I., Oppenheim, D., Buchsbaum, H., Emde, R.N. & The MacArthur Narrative
Group (1990). MacArthur Story Stem Battery. Unpublished manuscript.
Bretherton,,I., Prentiss, C. & Ridgeway, D. (1990). Family relationships as represented in

fifty-four
task
thirty-seven
at
and
a story-completion
months of age. In 1. Bretherton
Children's
M.
W.
Watson
(Eds.
),
Perspectives
Family
(pp.
85-106).
San
the
and
on
Francisco: Jossey-Bass.
Bretherton, I.,,Ridgeway, D. & Cassidy, J. (1990). Assessing internal working models of the
attachment rejationship: An attachment story completion task for 3-year-olds. In
M. Greenberg, D. Cicchetti and E.M. Cummings (Eds.), Attachment in the preschool
years
(pp. 273-308). Chicago: Chicago University Press.
Buchsbaum, H. K. & Emde, R.N. (1990). Play narratives in 36-month-old children: Early
family
development
Psychoanalytic
Study
Child,
45,
the
and
relationships.
moral
of
129-155.
BLichsbauni,H. K., Toth, S.L., Clyman, R. B., Cicchetti, D. & Emde, R.N. (1992). The use
of a narrative story stem technique with maltreated children: Implications for theory
and practice.
Development and Psychopathology, 4,603-625.
282
Campbell, S.B. (1995). Behaviour problems in preschool children: A review of recent
36
Disciplines,
(1),
Allied
Journal
research.
ofChila`Psycholo5) andPsychiatry and
113-149.
Cassidy, J. (1994). Emotion regulation: Influences of attachment relationships. In Nathan

A. Fox (Ed. ), The development of emotion regulation: Biological and behavioural
Child
Societyfor
Research
Development,
Monograph
59
in
the
of
considerations.
(2-3, Serial No. 240).
Cassidy,J. & Marvin, R. (1989/92). Attachment organisation in three- and four-year-olds:

Coding guidelines (unpublished manuscript). University of Washington, Seattle:
MacArthur Attachment Working Group.
Chodorow, N. (1978). The reproduction of mothering; Psychoanalysis and the sociology

Press.
California
University
Berkeley:
of
ofgender.
Occhetti, D., Cummings, E.M., Greenberg,M. T. & Marvin, R. S. (1990). An organisational

beyond
infancy:
for
Implications
theory, measurement,
perspective on attachment
Cicchetti
Curnmings
Greenberg,
D.
E.
M.
(Eds.
),
In
M.
T.
Attachment
and
and research.
Theory,
University
intervention,
Chicago
the
research
and
of
preschool years:
I.n
Press: London.
Crittenden, P.M. (1992). Quality of attachmentin the preschoolyears. Developmentand

PsYchopathologri,4 (2), 209-241.
283
Crittenden, P.M. (1997). Patternsof attachment and sexual behaviour: Risk of dysfunction
Zucker
(Eds.
J.
).
K.
integration.
In
L.
Atkinson
for
and
versus opportunity
creative
Attachment and Psychopathology, London: The Guilford Press.
Diamond, D. & Blatt,, S.J. (1994). Internal working models and the representational world
in attachment and psychoanalytic theories. In M. Sperling and W. H. Berman (Eds.),
Attachments in adults: Theory assessmentand treatment, pp. 72- 97.
Dodge, K., Pettit, G. & Bates, J. (1994).
Socialization mediators of the relationship
between socioeconomic status and child conduct problems. Child Development, 65,
1025-1034.
Eagle, M. (1995). The developmental perspectives of attachment and psychoanalytic theory.

In S.Goldberg, R.Mulr and J.Kerr (Eds.), Attachment theory: Social, developmental,
Press.
The
Analytic
London:
Fnide, R.N. (199 1). Positive emotions for psychoanalytic theory: Surprises from infancy
Psychoanalytic
American
directions.
Journal
Association,
the
of
new
researchand
39 (supplement), 5-44.
Fooel, G. (1989). The authentic function of psychoanalytic theory: An overview of the

The
Psychoanalytic
Loewald.
Quarterly,
Hans
58
(3),
419-451.
of
contributions
FonagN,P., Steele. H. & Steele, M. (199 1). Maternal representations of attachment during
284
pregnancy predict the organisation of infant-mother attachment at one year of age.
Child
Development, 62,891-905.
Fonagy, P., Steele, M., Moran, G., Steele, H. & Higget, A. (199 1). Measuring the ghost in
the nursery: A summary of the main findings of the Anna Freud Centre - University
College London Parent-Child Study. The Bulletin of the Anna Freud Centre, 14 (2),
115-131.
Fonagy, P., Steele, M., Moran, G., Steele, H. & Higget, A. (1991). The capacity for
in
The
reflective self parent and child and its
understanding mental states:
for
Infant
Mental
Health
Journal,
13,200-216.
of
attachment.
security
significance
Freud, S. (1895). Project for a scientific psychology. In J.Strachey (Gen. Ed.), The standard
Sigmund
Freud
(Vol.
1).
London:
the
of
complete psychological works
edition of
Hogarth.
Freud, S. (1914). On narcissism: An introduction. In J.Strachey (Gen. Ed. ), The standard

Freud
(Vol.
XIV).
London:
the
ofSigmund
completepsychological works
edition of
Hogarth.
Freud, S. (1914). Instincts and their vicissitudes. In J.Strachey (Gen. Ed. ), The standard
Freud
(Vol.
XIV).
London:
works
ofSigmund
psychological
complete
ofthe
edition
Hooarth.
r-I
285
Freud, S. (192 1). Group psychology and the analysis of the ego. In J.Strachey (Gen. Ed. ),
The standard edition of the complete psychological works ofSigmund Freud (Vol.
XVIII).
London: Hogarth.
Freud, S. (1926). Inhibitions, symptoms and anxiety. In J.Strachey (Gen. Ed.), The standard
Sigmund
Freud (Vol. XX). London:
the
edition of
completepsychological works of
Hogarth.
Freud, S. (1933). The new introductory lectures on psychanalysis. In J.Strachey (Gen. Ed.),
The standard edition of the complete psychological works of Sigmund Freud (Vol.
XXII). London: Hogarth.
Fury, G., Carlson, E.A. & Sroufe, L. A. (1997). Children's representations of attachment
in
family
drawings. Child Development, 68 (6), 1154-1164.
relationships
Gardner, H. (1987). The mind's new science. New York: Basic Books.
George, C., Kaplan, N. & Main, M. (1996). Adult Attachment Interview. Unpublished
Department
Psychology,
University of California, Berkeley (third
manuscript,
of
edition).
Greenberg, M. T., DeKlyen. M., Speltz, M. L. & Endriga, M. C. (1997). The role of
g
in
attachment processes externalizing psychopathology in young children. In L.
Atkinson and K. J. Zucker (Eds.), Attachment and psychopathology (pp. 196-222).
.
286
The Guildford Press: London.
Greene, J. & D'Ofiveira, M. (1992). Learning to use statistical tests in psychology. Open
University Press: Milton Keynes.
Grossman, K. E. (1995). The evolution and history of attachment research and theory. In
S.Goldberg, R. Muir and J.Kerr (Eds.), Attachment theory: Social, developmental,
London:
The Analytic Press.
Grossman,K. E. & Grossman, K. (199 1). Attachment quality as an organiser of emotional

in
longitudinal
behavioural
and
responses a
perspective. In C. Parkes, J. StevensonHinde and P. Marris (Eds.), Attachment across the life cycle, 93-114. London:
Tavistock.
Grusec, J.E. (1997). A history of research on parenting strategies and children's
In
J.
E.
Grusec
internalization of values.
and L. Kuczynski (Eds.), Parenting and
3-22.
Wiley & Sons: New York.
internalization
children's
of values,
fludson, J.A., Gebelt, J., Haviland, J. & Bentivegna, C.
(1992). Emotion and narrative
in
structure
young children's personal accounts. Journal of Narrative and Life
HistorT, 2(2), 129-150.
Kagan, J. (1984). The miture of the child. New York: Basic Books.
287
Karm i loff-Smith, A. (1992). Beyond modularity A developmental perspective on cognitive
science. London :The MIT Press.
Kaywin, R. (1993). The theoretical contributions of Hans Loewald. Thepsychoanalytic study

of the child, 48,99-114.
Kinnear, P.R. & Gray, C.D. (1994). SPSSFor Windows made simple. Lawrence ErIbaum:
Hove.
Kline, P. (1993). The handbook ofpsychological testing. London: Routledge.
LaPlanche. J. & Pontalis, J.-B. (1988). The language ofpsychoanalysis.

)
London: Karnac
Books.
Levin, H. & Wardwell, E. (1962). The researchuse of doll play. Psychological Bulletin, 59
(1), 27-56.
Mac Donald, K.
(1997). The coherence of individual development: An evolutionary
internalisation
perspective on children's
of parental values.
In J.E. Grusec &
L. Kuczynski (Eds.), Parenting and Children's Internalization of Values. New

York: John Wiley and Sons, Inc.
Nlain, M. (1994). A move to the level of representation in the study of attachment

for
Implications
Paper
for
psychoanalYsis.
the Annual
presented
organisation.
288
Research Lecture to the British Psycho-Analytical Society : London, July 6,1994
Main, M. & Cassidy, J. (1988). Categories of response to reunion with the parent at age 6:
from
infant attachment classifications and stable over aI month period.
predictable
Developmental Psychology, 24,415-426.
Main, M., Kaplan, N. & Cassidy, J. (1985). Security in infancy, childhood and adulthood:
level
to
the
of representation.
move
In 1. Bretherton and E. Waters (Eds.),
Growing points ofattachment theory and research. Monographs of the Societyfor

Research in Child Development 50 (1-2, Serial No. 209), pp 66-104.
Marans, S., Mayes, L., Cicchetti, D., Dahl, K., Marens, W. & Cohen, D. (199 1). The childfor
interview:
A
technique
studying thematic content. Journal
psychoanalytic play
Psychoanalytical
Association,
39(4),
1015American
1036.
the
of
Marans, S., Mayes, L. C. & Colonna, A. B. (1993). Psychoanalytic views of children's play.
In A. Solnit,
D. Cohen and P.Neubauer (Eds.), The many meanings of play. - a
9-28),
(pp.
psychoanalytic perspective
NLt\,seless,0. (1998). Maternal caregiving strategy -A distinction between the ambivalent

Infant
Mental
Health
Journal,
disorganised
19(l),
20-33.
the
profile.
and
&
Carlson,
Kalkoske,
M.
E.
B.,
A.
(1994).
Egeland.
Relationships
R.,
E.
between
IMcCi-mic,
carly maltreatment and mental representations of relationships assessed with
289
6(l),
Psychopathology,
Development
in
and
projective storytelling middle childhood.
99-120.
Moore, B& Fine, B. (1990). Psychoanalytic terms and concepts. New Haven: American
Psychological Association and Yale University Press.
Oppenheim, D., Emde, R.N. & Warren, S. (1997). Children's narrative representations of
development
Their
and associations with child and mother adaptation.
mother's:
Child
Development, 68(l), 127-138.
Oppenheim, D., Emde, R.N. & Winfrey, N. L. (1993). Aggression and coherence in
family
Links
Paper
with
relationships and adaptation.
children's narratives:
Construction
and co-construction of pre-schoolers'
presentedas part of symposium
family
influences,
for
Society
Research
in
Child
narratives about emotions:
Development, March 1993, New Orleans.
Oppenheim, D., Emde, R.N., Warnbolt, F.S. (1996). Associations between 3-year-olds
fathers
and their story completions about
narrative co-constructions with mothers and
Early
Development
Parenting,
5
(3),
149-160).
themes.
and
affective
Oppenheim, D., Nir, A., Warren, S. & Emde, R.N. (1997). Emotion regulation in mother
Associations
e co-construction:
with children's narratives and
child nairratiN,
Psychology,
Developmental
(20),
284294.
adaptation.
-333)
290
Oppenheim, D& Waters, H. S. (1995). Narrative processesand attachment representations:
Issues of development and assessment. In E. Waters, B. Vaughn,
G. Posada and
K. Kondo-lkemura (Eds.), New growing points of attachment. Monograph of

Child
Development,
60
(2-3,
Serial
Research
No. 244).
Socieryfor
in
the
Raven, J.C., Court, C. & Raven, J. (1986). Manualfor Raven's progressive matrices and
London:
H.
K.
Lewis
&
Co.
Ltd.
scales.
vocabulary
Raynor, E. (1991). The independent mind in British psychoanalysis. London: Free

Association
Books.
Renfrew, C.E. (199 1). (Second Edition). The Bus Story Test -A test of continuous speech.
Old Headington: Oxford.
Robinson, J., Mantz-Simmons, L., Macfie, J. & The MacArthur Narrative Working Group
(1992). The narrative coding manual. Unpublished manuscript.
Sandier, J. & Rosenblatt,, B. (1962). The concept of the representational world. The
PsYchoanalytic study of the child, 17,128-145.
Siegel, S. & Castellan, N. J. (1988). Nonparametric statistics for the behavioural sciences
(2nd edition). London: McGraw-Hill.
Slade, A. (1993). A.Iftct regulation and defence: Clinical and theoretical considerations.
291
Unpublished discussion presented as part of symposium on affect Regulation and
Attachment at the Biennial Meetings of the Society for Research in Child
Development, New Orleans, LA.
Slade, A. (1997). Making meaning and making believe: Their role in the clinical process.
In A. Slade and D. P. Wolf (Eds.), Children at play. - Clinical and developmental
81-107).
(pp.
to
meaning
and
representation
approaches
New York: Oxford
University Press.
Slade, A. & Aber, L. (1992). Attachments, drives and development: Conflicts and
in
In
J.
Barron,
M.
Eagle
D.
Wolitsky
(Eds.
),
Interface
theory.
and
convergences
(pp.
154185).
Washington,
D.
C.
American
:
ofpsychoanalysis andpsychology
Psychological Association.
Solnit, A. J. (1987). A psychoanalytic view of play. Psychoanalytic study of the child, 42,
205-219.
SPSSBase 7.5for Windows.- Users Guide. (1997) Chicago: SPSS Inc.
Speltz,M. L., Greenberg,M. T. & de Klyen, M. C. (1990). Attachmentin preschoolerswith

disruptive behaviour: A comparison of clinic-referred children and non-problem
2,31-46.
Devc1opment
psychopathology,
and
children.
SroutIe, L. A.
(1986).
Appraisal: Bowlby's contribution to psychoanalytic theory and
292
developmental psychology: Attachment: Separation: Loss.
Journal of Child
Psychology and Psychiatry, 27 (6), 841- 849.
Stem, D. (1985). The interpersonal world of the infant. New York: Basic Books.
Steele, H. (199 1). Adult personality characteristics andfamily relationships: The

development and validation of an interview based assessment.Unpublished Ph.D
thesis, University of London.
Steele, M. (1990). Intergenerational patterns of attachment. Unpublished Ph.D. thesisq

University of London.
Steele,H., Steele, M. & Fonagy, P. (1996). Associations among attachment classifications

fathers
infants.
development,
Child
67
(2), 541-555.
their
and
of mothers,
Tarullo, L. B. (1994). Windows on social worlds: Gender differences in children's play

Children
D.
P.
Wolf
(Eds.
),
Slade
Play,
Oxford
In
A.
University
and
at
narratives.
Press: Oxford.
I'lionias, A. & Chess, S. (1977). Temperamentand development.New York: Bruner/Mazel.
"I'lionipson, R.A. (1994). Emotion regulation: A theme in search of a definition.
In
Nathan A. Fox (Ed.), The developmentof emotion regulation: Biological and

behavioural considerations. Alonograph oj'the Societyfor Research in Child
293
Development, 59 (2-3, Serial No. 240).
Furner, P.J. (199 1). Relations between attachment, gender, and behaviour with peers in
Child
Development,
62
(6),
1475-1488.
preschool.
&
Emde,
R.
The
Structure
K.,
Kelsay,
K.
N.
(1998).
Five-Year-Old
Klitzing,
of
von
Children's Play Narratives Within the MacArthur Story Stem Methodology.
Unpublished Manuscript, University of Colorado Health SciencesCentre, Colorado.
Warren, S.L., Oppenheim, D. & Emde, R.N. (1996). Can emotions and themes in children's
behaviour
?
Journal
American
Academy
Child
the
problems
of
play predict
of
and
Adolescent Psychiatry, 34 (10), 1331 - 1337.
Winnicott, D. (1965).
The maturational processes and the facilitating
environment.
London: Hogarth.
\koolgar, M. J. (1996). Parental, infant and early childhood attachment classification and
its relationship
to sociomoral development.
Unpublished Doctoral Thesis,
University College London.

Loewald, H. (1978). Instinct Theory, Object Relations
and Psychic Structure
Formation. Journal of the American Psychoanalytic Association,
26 (3), 493506.
Piapet, J. (197)).
L
The Stages of the Intellectual Development

of the Child. In S.
Harrison& J. McDermott (Eds.). Childhood Psychopathology.

New York:
294
APPENDIX A. 1
MacArthur
I.
Story Stem Battery Protocol
Story Stem:
SPILLED JUICE
Characters:
Mother, father, child (Susan/George) and younger same sex sibling

(Jane/Bob)
Table, pitcher
Props:
'The family is thirsty and they are going to have somejuice. Now, put
Interviewer:
the family around the table so they can have some juice'. (wait until the figures are placed
by the child or interviewer) 'Here's the family drinking their juice. Susan (or George if
her/his
is
Uh,
Oh
juice
the
table
and,
s/he
spilled
subject male) gets up and reachesacross
,
doll
knock
floor
floor
it
is
!'
(make
the
the
that
the
child
pitcher onto
so
visible to
all over
the subject)
Interviewer:
Now you show me and tell me what happens next
Prompts:
(If nothing is done about the juice)
'What happens about Susan/George spilling the juice ?
'Who cleans up T
'Does anyone say anything T
'How do mum and dad feel about Susan/George spilling the juice T
2.
Story Stem:
MOTHER'S HEADACHE
Characters:
Props:
Mother, child, samesex friend

Couch, chair, television
Interviewer:
'Mom and Susan/George are sitting on the couch watching
television.' (mom turns to child) 'Oh Susan/George,I have such a headache !I just have to
turn the television off and lie down. (mom gets up and turns television off) 'Susan/George,
find
do
for
?'
(Susan/George respond) 'Okay mom. I'll
to
can you
something quiet
a while
book.
lies
down
'
Susan/George
(Mom
in
the
read my
on
couch and
sits the chair to read a
book) (Interviewer makes a doorbell sound and moves child figure to answer the door) 'It's
Susan/George'sfriend, Laura/Dave.' Laura/Dave say 'There's this really neat TV show on.
Can I can in and watch it with you ?'
Interviewer:
Now you show and me and tell me what happens next?
Prompts:
/Georoe don't turn on the televisio

(If Susan.
I n)
Law-a/Davesay. 'Oh come on !I know you'll really like it. '
295
APPENDIX A. I
(If any child turns on the television)
Mother says 'Oh, I have such a headache P
I
Story Stem:
THREE'S A CROWD
Characters:
Props:
Child, samesex younger sibling, samesex friend, mother, father

ball
Interviewer: Mom and dad are over here talking to the neighbours. Susan/George
is playing with her/his friend, Laura/Dave and her/his new ball. Show me how they play
little
Susan's/George's
ball.
(Allow
to
the
sister/brother comes
child
play a moment)
with
Susan/George
house
I
T
'Sure'.
But
'Can
the
says,
of
and
says,
play
with
you
running out
Laura/Dave says, 'No way ! If you let your little sister/brother play, I won't be your friend
!'
anymore
Interviewer:
Show me and tell me what happens next ?
Prompts:
(if request is ignored)
'What about Jane/Bob ? She/he wants to play with them ?
(If Susan/George does not come to sibling's defence)
Sibling says 'But Susan/George,I'm your little sister/brother P
(If Jane/Bob is immediately allowed to play by Susan/George)
Laura/Dave say, 'But I said I didn't want to play with your sister/brother. I'm
leaving !'
Story Stem:
BURNED HAND
Characters:
Props:
Mother, father. child, younger sibling

)
Stove, pan, table, chairs
Interviewer: Mom and Susan/Georgeare at the stove. Dad and Jane/Bob are sitting
have
but
it's
Mom
to
6We're
the
table.
a really good supper
going
at
says,
not ready yet.
I)oii't get too close to the stove.' Susan/Georgesays, 'That looks good. I don't want to wait.
I'd like some now !' (Susan/George knocks the pot off the stove) 'Ow
I've burnt my
hand ! It hurts !'
Interviewer:
Now you show me and tell me what happens next ?
Prompts:
(if no one helps the child)
What about Susan/George's burnt hand ?
(if no response)
296
CC-On-t-)
A-1
APPENDIX
What do they do about the burnt hand ?
(If bum is attendedto)

Do they say anything about the hurt hand ?
5.
Story Stem:
LOST KEYS
Characters:
Props:
Mother, father, child

none
Interviewer:
Susan/Georgecome into the room and seemom and dad looking at
in
face
for
(mother
face.
(make
like
Look
the
this.
subject)
angry
an angry
at my
eachother
did
did
P
(mother)
'Yes
keys
!'
(father
'I
lost
'You
always
you
you
angrily)
not
my
voice)
,
lose my keys !' (father) 'Well, I didn't lose them this time P
Interviewer:
Prompts:
(If no response or argument is ignored)
What happens about mom and dad's argument ?
6.
Story Stem:
SWEET SHOP
Characters:
Props:
Shopkeeper, mother, child

Counter or shelf
'Here we have the sweet shop and the shopkeeper. Do you know
Interviewer:
Susan/George.
Here
Candy/sweets.
Susan/George
?
the
comes mom and
what's on
shelf
?'
had
have
Mom
'No,
!
May
I
'Oh,
says,
you already
some
some today. Let's
says,
sweets
from
Susan/George
home.
'
(Mother
takes
the shelf) The shopkeeper
a sweet
go
walks away.
?'
doing
(Mom
turns around)
there
'Hey,
says,
what are you
Interviewer:
Prompts:
(if no response to the stealing)
Susan/George took a candy/sweet. What does mom say about that ?
(if mom responds to stealing)
Does Susan/George say or do anything ?
7.
Ston, Stems: SEPARATION
Characters:
Props:
Mother, father, child, sibling. grandmother

car
297
APPENDIX
Interviewer:
A. 1
'You know what it looks like to me ? It looks like mom and dad are
house.
Mom
in
front
(bring
is
The
*OK
the
trip.
says.
out
car)
of
parked
car
on
a
going
Grandma
We'll
dad
I
tomorrow.
trip
see
will
now.
you
and are going on our
girls/boys, your
'
stay with you.
The interviewer should let the subject put the figures in the car and
Important:
if
drive
intervene
Only
drive
the
to
the
them
subject seemsunable make
car
off
off
make
If the subjectputs the children in the car say, 'No, only the mom and dad are going.' After
the subject (or interviewer if necessary)makesthe car drive off, then interviewer puts the
interviewer
If
the
to
the
'No,
table
the
subject
of
sight.
wants
retrieve
car
says,
out
under
car
they're not coming back yet.'
Interviewer:
Prompts:
What do the girls/boys do now that mom and dad are gone ?
8.
Story Stem:
REUNION
Characters:
Props:
Grandmother, child, sibling, mother, father

car
Interviewer: (in monotone voice) It's the next day and Grandma looks out the
dad
home
from
look
I
'Oh,
think
their
mom
and
girls/boys,
your
are
window and shesays,
trip. I think I can see their car.'
Bring the car with the two parents back out from under the table and
Important:
keep
it
interviewer
from
distance
le.,
has
the
the
the
to
children.
near
so
set it at a
subject
it
drive
for
it
'home'
reach
and make
Interviewer: Now you show me and tell me what happens next ?
What do they do now that mum and dad are home ?
Prompts:
Story Stem:
BATHROOM SHELF
Characters:
Props:
Mother, child, younger sibling

Bathroom shelf, bathtub, toilet
Interviewei-: 'Now it's the next morning and dad has gone on an errand. (Bring out
for
?
The
Can
in
these
the props)
are
their room
room
what
girls/boys
are
guess
playing
you
in
Girls/boys,
Mum
have
I
here
door
*
toys.
to
their
and
says,
comes
to the
o\-cigo next
with
be
but
back.
I'll
Don't
bathroom
touch
right
the
to
something,
anything
return
on
iieiAbours
tN
'
Mum
"boys
'Okay
(put
The
doll
the
to
mum.
say.
goes
neighbours
oirls,
mother
shelf, okay"
JaneiBob
Jane/Bob
Susan/George
Jumps
Ow
then
and
play
some
more
table)
the
up!
under
-
298
APPENDIX
A. 1
!I cut my finger. I need a plaster !' Susan/George says, 'But mum said not to touch
finger,
it's
bleeding
!'
bathroom
'But
Jane/Bob:
the
my
shelf'
anything on
Interviewer:
Now you show me and tell me what happens next
Prompts:
(If any child gets a plaster)
Mother returns and says, 'Hi, kids, I'm back. '
(If there is no mention of the plaster)
Mother says, 'Jane/Bob, I see a plaster on your finger. I thought I told you not to
touch anything on the bathroom shelf. '
(If child does not get a plaster)
Jane/Bob says again, 'My finger's bleeding
10.
Story Stem:
EXCLUSION
Characters:
Props:
Mother, father, child

Couch
Interviewer: Mum and dad are sitting on the couch talking. If subject is female,
father
is
George
Susan/if
turns
to
turns
to
and says,'Mom/dad and I
subject male
mother
like
Will
time
some
alone.
you pleasego up to your room and play with your toys.
would
Pleaseshutthe door so it is quiet. (Allow the subjectto move Susan/Georgeaway) After the
hug.
Susan/George
then
mom/dad gives mom/dad a
subjectmoves
Interviewer:
Prompts:
(If the child goes to his parents)
Mum/dad says, 'We asked to have some time alone.'
(If child complies with request)
Mum/dad says, 'Okay, Susan/George.Thanks for letting us have some time alone. )
11.
Story Stem:
BISCUIT TfN
Characters:
Props:
Mother, father, child and younger sibling

Table, biscuit tin
Interviewer:
Susan/George and Jane/Bob are in the kitchen. Jane/Bob sees the
Susan/George
biscuits
biscuit.
NO
!'
Jane/Bob
bISCLIlt
'Mum
takes
tin and
says,
said
a
says,
know
it!
dad
'
You
don't
tell mum and
about
'Please
what ? HERE COMES MUM AND
1),M) ! (\\,Ith emotion in voice)
299
APPENDIX
Interviewer:
A. 1
Prompts:
(If nothing was said about the biscuit being taken)
Mum/dad says, 'I see someone took a biscuit. Who was it T
300
APPENDIX A. 2
MACARTHUR NARRATIVE CODING MANUAL
JoAnn Robinson, Linda Mantz-Simmons, Jenny Macfie
and the
MacArthur Narrative Working Group
Note:Additions to the text in this typeface are the results of discussion between the
Linda
Mantz-Simmons during the training course at University
team
and
coding
College, London.
INTRODUCTION
Coding for each of the narrative stemshas been divided into 3 sequentialphases,
development,
the
the
presentation, narrative
namely
and the transition betweennarratives
beginning
The
has
been
defined
follows:
and end of eachphase
phases.
as
1) Presentation Phase - begins once the examiner begins to deliver the first line of
the narrative script. This phase ends once the examiner has presented all the details of the
has
has
story,
physically releasedall props, and
prompted the child to finish the story for the
first time (ie. using the statement "Show and tell me what happens next").
If the subject picks up a doll or prop during the presentation phase but does not actively
do
it,
however,
if
in
the
the subject should attempt to begin the
engageit
narrative,
not code
beginning
i.
this
the
the
narrative, record
as
of
narrative phase. e the presentation and
narrative phases may overlap.
2) Narrative Development Phase - begins once the child emits their first response
towards the story I ine. If the child begins the narrative just before the examiner states" Show
me and tell me what happens now" and the examiner never delivers this line or delivers it
beginning
begun,
has
first
A
this
the
the
the
after
consider
of
child
narrative.
responsewould
include such things as the child asking questions or picking up the dolls, this would not
just
the
include
staring at the table, which may be indicators of
child walking away or
avoidance.This phase ends one of 2 ways, a) once the child makes a clear statement about
the story line being over (ie. they all go to bed), or b) once the examiner indicates that the
have
"I
is
(ie.
story over either verbally
another story") and/or physically (ie. removes the
first prop). Choose between these basedon the presenceor absenceof the first choice. If the
for
to
ask
more of the story after the child has clearly finished and the
examiner continues
in
domain
disorganizes
then
or makes shifts
child
stop coding at the point at which the child
for
the
and
only
code
responsivity with the examiner until the examiner
completed
narrative
Think
how
long
the
has
into
been invested in their
of
transition
the
child
phase.
moves
story.
3) Transition Between Narratives Phase - begins one of 2 ways, a) once the child
has made a clear statement about the story being over, or b) once the examiner has indicated
This
begins
delivery of the first
the
ends
once
phase
examiner
that the story is over.
301
APPENDIX
A. 2
Once
has
finished
last
for
following
the
the
the
narrative.
child
story
charactersstatement
for
Transition
finished
two
though
the
them
a
phaseeven
a minute or
are
in the series,give
the narratives,thus providing them with the opportunity to add to or changetheir previous
during
be
in
Themes
the
transition
that
phaseshould noted the comment
occur
narrative.
section.
The Family Birthday narrative is not coded as it is used as an introduction to the testing
by
indicating
is
being
Begin
coding
which story stem
coded and whether
proceduresonly.
it's being done in the home or lab. Proceed to fill in the start time for the first phase, indicate
it
the presenceof each content code as appears, rate each performance section that pertains
to the phaseyou are watching, and note the end time for that phase. The total times for each
finished
be
Proceed
for
through
these
once
are
coding.
calculated
you
can
steps
each
phase
The
Over-All
(OA)
is
the
the
to reflect the
narratives.
column
on
coding
of
sheet
phase
behaviour
3
the
over-all
narrative phases.
subjects
Examiner errors may occur in several ways which will effect how the narrative is to be
key
in
1)
During
the
the story may be deleted (ie. in the
presentation phase a
point
scored.
don't
kiss).
do
In
the
this
parents
case not code the narrative and write a note
exclusion story
2)
During
development
to
the
the
comments
as
why
narrative
was
not
coded.
narrative
under
incorrectly
deliver
the
a prompt or they may add a prompt that changes
phase examiner may
the story line. If the child has had time to develop a portion of the narrative, code everything
inadvertently
Note
in
the
the
to
the
this
the
at
examiner
changes
story.
up
point
which
had
before
has
begin
If
the
this
time
to
occurs
child
comment section.
much of their
in
do
this
the comment section. 3) In some
the
narrative not code
narrative section and note
has
long
too
the
the
after
child
completed their story to end the
situations
examiner may wait
fill
in
the
to
try
that time with disorganized or
this
child
and
narrativephaseand
may cause
incoherentstatements.When this occurs, code only the coherent portion of the child's story,
record the time at which you stop coding as the end of the narrative phase and note this in
the comment section.
Authors David Oppenheim and Sun Park deserveseparaterecognition for their contribution
in
Parental
Representations
Narratives".
"Coding
the
of
of
section entitled
1.CONTENT THEMES - coding will be basedon the presenceor absenceof the following
themes through-out the presentation, narrative development and transition phases of each
it
box
Place
to
theme
the
each
as
story.
occurs only once per narrative.
corresponding
a/ in
Themes that occur during the Presentation or Transition phases that are not related to the
These
Comments
be
be
themes
the
section.
iiarrative may
may
noted in
stated and/or
eiiacted by the subject.
A) NO THEMES:
there are no themes evident through-out the narrative.
302
APPENDIX
B) RELATIONSHIP
A. 2 (cont)
THEMES:
1) COMPETITION (CM) -a dyadic relationship between children striving for

include
This
the sameobject or activity.
may
competitive comments, complaints about turns
This
does
include
fairness,
to
not
competition
comparisons
of
other
self.
or negative
or
Largely
dyadic
between
the
attention.
verbal,
and
parents
children,
over/for
Ex: - "No, I got it"
better
"But
I
than
can
ride
you"
"They're
"
children
argue
over
cookies
mine!
"You
!"
can't
catch
me
2) SHARING (SH) - this denotes a positive relationship between children
include
interactions
These
There
be
the
sharing
an
object
of
or
animal.
must
and/or adults.
doll
If
level
up. a
offers to share without giving
of a character giving
some
-something
it
anything up code under affiliation.
Ex: - child offers the bike or horseto their sibling to ride

in
it
hang
together
to
picture
order
share
parents
up
(R/J) -a triadic relationship between children over
3) RIVALRY/JEALOUSY
between
children alone.
adult attention, or
Ex: - child goes over to father who is holding brother and asks to be picked
dad
"He's
too"
my
up while exclaiming
4) EXCLUSION (EX) I= Other - character prevents another from joining in an activity or a
fits
if
it
too)
(Check
punishment
character gets sent away.
Ex: - parent sending a child to their room
bike
but
doll
the
to
getting ride on
one
- all
Self - character excludes self from activity or others, to include isolation
from
and withdrawal
conflict.
Ex: - "Exclusion" child says she doesn't want to go into room with
hurt.
parents as she'll get
Both I and 2
Unless it is clear why the child is going away, do not code. You must
have the sense that someone is being sent away, not wanted.
303
APPENDIX
A. 2 (cont)
5) AGGR-ESSION (AG) I= Physical - aggressiondirected by the subject or characterstowards another

interactions
These
have
include
to
them
or
object.
a
negative
prop,
quality
and
character,
hostile, destructive gesturesand forms of physical aggressionsuch as an object being thrown
intent
This
does
include
the
to
with
cause
character
pain.
category
at another
not
angry
intended
discipline.
that
acts
are
as
punishment
physical
or
or
yelling
Ex: - subject has one doll hit or punch another doll
doll
horse
bike
the
one
pushes
another
off
or
doll
hits
horse
2= Verbal - aggressivecomments, excluding shaming, blaming, comments.
Ex: - "I hate you"
Including threats without a disciplinary motive.
3= Both
6) EMPATHY/HELPING
(E/H) -a character or the subject either identifies
demonstrates
feelings
the
thoughts
an understanding of
or
with or
of another. This may be
demonstratedby a worried or concerned facial expression or tone of voice, a movement or
gesture toward or visually attending to the victim. Also, the character or subject seeking
from
divert
the
to
the victims attention,
reassurance
victim, mother or examiner, attempts
helping
by
the
the
victim
sharing something with
performing an act to relieve
victim, or
distress.
Helping behaviours would include one doll helping another to perfon-n a task or providing
faster.
does
include
job
done
This
doll
doing
that
not
assistanceso
correctly or
one
a
gets
an
independently
juice
for
Mom
the
act
cleaning up
or the subject cleaning up
another such as
the room.
first
in
described
the
type
as
of
empathy
empathy/psychic above
Ex: - one doll talking about the injured party to another doll,
describing the owie or the hurt
doll
injured
toy
to
the
offers
a
party
one
"I
hurt
knee
too"
saying
my
once
subject
Story "No, because my mom's got a
Headache
Mom'
In
headache" Just saying "No, we can't" is not enough.
2= helping behaviour - as in the second type described above
Ex:- mom doll assisting child doll in wiping up the juice
both
304
APPENDIX
A. 2 (cont)
(REH) - one character approachesanother

7) REFUSED EMPATHY/HELPING
for empathy/helping and it is actively denied. Must be quite explicit.
Ex: - child asks mom to help find Barney and mom says "No"
8) INTERPERSONAL CONFLICT RESOLUTION (ICR) - describessituations
in which the subject has set up or identified an interpersonal conflict between the dolls, and
in
This
be
to
the
through
successfully
resolve
conflict.
which
means
may
various
uses
final
This
decision
to
third
the
party act as a mediator or as
maker.
sharing, seeking out a
between
intra
be
This
characters
and
not
an
psychic
conflict.
code
conflict must
between
to
characters.
conflicts of needs
must refer
1. =. Seekshelp in conflict resolution from a 3rd party for self gain, not just to get
look
better
doll
in
(code
trouble
to
self
or
make
as tattling).
another
Ex: - one child wants to ride the other child's bike, the one in possession of
the bike does not want to share so the child v-oesto mother and tells her the 12roblem,mother
both
have
turns,
take
to
they
children comply.
and
says
Three's
Crowd
in
1
by
If
is
the
a
code
child
a
parent
used
2. = Child uses an "adult" like strategy to resolve conflict
Ex: - Crowd story - children negotiate a time limit between each other so that
everyone gets a turn.
TV
house
Headache
Mum's
child's
at
other
watching
-
the ball"
3. = Threats are used to resolve conflict, also include threats to tattle

Ex: - 3's crowd story - "If you don't let my brother play I won't give you
it"
it
have
I'll
I
"If
ruin
can't
I'll
hit
"Don't
Tin
tell
Biscuit
or
you"
story
Check to see if the threat involves aggression; if so code.
9) TATTLING (TAT) - one doll "tattles" on another in order to get the other
in trouble or to make themselves look better. Do not include instances where the motive is
blame).
Threatening
is
(CAUTION
(ICR-1).
to
tattle
self inclusion or gain
-not same as
coded under ICR-3.
Ex: - "Mom, I saw him steal the cookie"
in
(CP)
COMPLIANCE
to
to
the
10)
situations
which
a
character
refers
vields
have
is
Child
demonstrated
does
asked.
must
what
some
rules or requests of an adult and
understanding of the stem to make compliance meaningful.
305
APPENDIX
A. 2
Ex: - child wipes up juice upon request from parent

- child cleans up room upon request
to
goes
room upon request
child
TV
turning
on
not
(N/C) - refers to situations in which the character
11) NON-COMPLIANCE
does
in
it.
The
the
to
and/or
something
or
request
rule
opposition
character may also
ignores
breaking
before
breaking
it.
Child must
they
a
rule
or
and
justify
while
are
acknowledge
havedemonstratedan understanding of the story stem to make non-compliance meaningful.
Ex: - child gets a band-aid from bathroom shelf while saying I know we're not
but
hurt"
Johnny's
to
supposed
before
cleaning room
- child plays with special present
in
Biscuit
Tin
do
biscuit
not
code
eating
as non-compliance
12) ESCALATION OF CONFLICT (E/C) - indicate incidences of a character
level
beyond
first
the
that
of
aggression
evident in either characters
escalating
expression of
include
Also
level
the
acts
of
retaliation
may
reflect
aggression.
which
same
of aggression
that was initially inflicted upon them. These may also be coded under the aggression
include
in
involve
discipline
Do
this
that
category situations
category.
not
or punishment in
Look
for
from
have
both
Mom
talking
to
when
you
sternly
spanking.
may go
which
verbal and/then physical aggression. Also include references escalating to
suggestions of divorce.
Ex: - Mom yells at child, child hits mom.
back
knocked
horse
hits
the
that
gets
off
aggressor
child
Lost Keys - child continues argument and then character becomes
aggressive
also can be verbal conflict, ex. An argument escalating and divorce
suggested
13) SHAME (SM)- character exhibits signs of embarrassment or may make

This
I
ftypes
shames
another.
one
character
usually
se reproaching
of statements, or when
hivolves angry vocs - with a harsh shaming tone. Also, check to see if
be
iscipline
coded too.
should
pun ishment/d
1. - Shame self - child says "I'm so stupid, I can't do anything right" after being
for
spilling juice
scolded
is
him
he
knows
Mom
hides
to
toys
catch
going
playing
vvith
when
child
iiistead of taking a nap
2. - Shame other - one character shaming another "Shame on you, you are bad"
Or "You are a nauglity girl"
done
"Look
you've
what
-
306
APPENDIX
A. 2
"You shouldn't have done that !"

Note
kiss
laughs
this
one under comments
s/he
see
parents
when
child
only
(AFL) - refers to situations in which all the dolls are
14) AFFILIATION
This
in
be
for
2
dolls
if
together.
there is
activity
code
an
made
used
or
more
participating
inclusion.
is
There
inclusion
belonging.
or
a
sense
of
or
a clear senseof affiliation
Ex: - everyone gets a turn on the bike
everyone goes to the park
In Three's a Crowd only if friend says "Let's all play
15) BLAME (BLM) - there are two categories of blame:

1. = Blame self -a doll blames self for act committed by others, may be viewed
doll
lost
ie.
I
keys"
Keys
the
towards
says
conflict reduction
as an effort
- child
have committed
Blame other - one doll blames another for an act that they may or may not
Must be in a direct way. Ex - "You did it! "

This includes accusatorycomments
Ex: - "What did you do with thosekeys?" (In a harshvoice)
3. = Both
(Caution - not same as tattling "She did it !" is tattling. )

16) AFFECTION
caring touch, or praise.
(AFC) - any display of hugs, kisses, compliments, wann or
Ex: - Mom telling child they did a good job

"Thank you"
"Good girl"
(T/T) - one character teasing another with an
17) TEASING/TAUNTING
defensive
form
is
Taunting
desire.
the
comment.
of a
often in
object that they
didn't
"Mom
Fx: - one child says to another
get one"
e me a cookie and you
(--Ya\,
didn't
I
have
if
I
don't
it
anyway"
want
it,
can't
care
-I
307
APPENDIX A. 2 (cont)
18) DISHONESTY
(DSH) -a character lies, steals or sneaks as part of the
narrative.
Ex: - doll hides cookie behind back and tells mom I don't have anything"
bandaid
bathroom
to
tiptoes
get
off
shelf
- cheating
keep
from
to
transgressions
(but
things
plotting
parents
not
good
like surprises)
19) PUNISHMENT/DISCIPLINE
(P/D) - The following forms of
discipline/punishment may be stated or enacted. P/D vocs don't have to be harsh. Include
threatsof P/D, or instances in which one child doll tell another that they are going to tell a
This
is
from
More
disciple
to
them.
the
that
parent
child
only.
of a
parent will
parentso
limit setting. You can score both pun ishment/discipline and aggression.
Verbal - punishment or discipline (or threat of) with no physical element
Ex: - time out, scolding, exclusion, deprivation, increaseddemands,rules
or po icies
2= Physical - punishment or discipline (or threat of) with a physical element.
'rhe subject or character inflicts pain or discomfort on another character which may take
disciplining
for
be
form
This
the
children, parents
parents
may
used
of spanking.
disciplining parents, children disciplining each other or the parents.
3= Both
4= Unspecified - child indicates that P/D takes place but they don't specify
the form it takes.
Ex: - "He gets in trouble"
in
his
(no
him
"Mom
vocs)
room"
puts
"Stealing"
"Mom takes him home" along with some indication that he's in
trouble
(R/G) - the act or process of a character making
20) REPARATION/GUILT
between
disharmony
following
feelings
the
displaying
child and/ or
some
guilt
aniends or
in
include
Also
be
this
This
category instances
or
non-verbal.
verbal
may
adult characters.
doll
makino things right again.
ot'die subjector a
reparation/practical
Ex: - fixing or righting the wrong
keys
finding
back
candy
puts
juice
the
gravy
or
up
cleaning
anyone
-
308
APPENDIX
A. 2 (cont)
2= reparation/guilt
Ex. - mom saying she's sorry and she's not mad anymore to the child
49
I'm sorry"
feel
they
sorry or naughty not sad
someone says
CONFLICT (VC) - includes highly inflected angU verbal

between
the
or
calling
yelling
children
and/or
adult characters.
as
name
such
remarks
21) VERBAL
Ex: - child yells back at mom "I'm not going to clean my room"
in
Lost
Keys
is
VC
the
argument
continuing
22) PERSONAL INJURY (PI) - pertains to any instance of a character being
be
inflicted
injury,
injured.
hurt
This
that
the
of
an
accident
result
may
a
self
or
physically
focus
be
injury
The
by
to
the
the
needs
on
or
child or a prop.
was caused another character,
injury
Subject
just
the
the
the pain, not
may present
story stem
on
act of aggression itself
in another character.
Ex: - child screamswhen failing off the rock (Failing off the rock or failing
down by itself is not enough)
injury
describes
the
- child
injury
described
hospital
to
with
going
illness
not
23) ATYPICAL RESPONSES (AR) - Write a brief description of these on the
line designated AR. (Make special note of death and killing responses under
comments)
I= Neutral or Positive - atypical responses that are not concerning or
alarming
Ex: - dad puts his head in his cup
examiiier with
2= Negative - any atypical or disorganized responses that leave the

has
bewilderment
that
tone
to
a
negative
concern
it.
or
a sense of
Ex: - "the house catches on fire and everyone dies"
crocodiles
-
3= Both I and 2 in samenarrative.

24) REPETITION
(REP) -
Same story - whenever the subject repeatsthe story line and or actions
it
be
be
This
the
the
by
same
or
may
exactly
the
may
repetition
examiner.
IListpresented
from
does
The
different
to
the
add
or
subtract
not
subject
character.
a
saine actions with
just
instances
the
the
Do
continues
parents argument
child
where
not include
orioiiial story.
because
"Keys"
this is a natural continuation of events.
as in
309
APPENDIX
A. 2
2= Prior story -a recurring theme such as the birthday party happening

is
intrusion
This
Barney"
doesn't
fit.
"Lost
theme
that
the
story.
an
the
of
a
end of
at
3= Both I and 2
Must get sense that the child is running out of things t say or is stalling, etc. If the
don't
in
a meaningful way,
code.
child uses repetition
CODING OF PARENTAL
REPRESENTATIONS
IN NARRATIVES
David Oppenheim, Sun Park, JoAnn Robinson,

Linda Mantz-Simmons, Jenny MaeFie
11.Theseare parent to child only, not parent to parent. When coding parental representations
look for moments in which parent is described as doing or saying something in the past,
if
future.
Also,
talks
the
about
parents even their actions are
notice when subject
presentor
describes
described,
the child-protagonist's expectations of the
the
subject
or when
not
her
his
Several
be
Do
to
the
or
actual parents.
codes can
child
not code referencesof
parent.
itself
it
is
if
for
However,
the
given only once.
samecode repeats
even
given
each narrative.
This strategyavoids getting into problems related to deciding on the boundaries of themes.
Code as follows:
I= Mother involved only
2= Father involved only
Both parentsinvolved
1) POSITIVE
(If the child says 'they' code 3)
(POS):
A) Protective: Parent is described asprotecting the child from possible or actual

harm.
Ex: - "Be careful with the scissors"

include
saving
include
character
of
child's
protective
-
B) Caretaking: Parent is described as engaging in caretaking actions. involving

teeding or taking care of child's hurt.
Ex: - parents put bandaid on finger, parent feeds the familv
310
APPENDIX A. 2
might be emphasis on feeding ; it is enough to say that mum
cooked dinner
parent takes child to hospital
Q Affectionate, warm, caring, supportive and affirming: a broad category for
descriptions:
Hugging, kissing, complementing child.
a range of positive
Ex: - "she likes to be with her Mom and Dad"
11giveMom and Dad a kiss"
Picture- emphasis on the child's' efforts or abilities.
D) Helpful: Parent is giving child concrete help or child seeks help from the
parent.
Ex: - parent helps child find lost dog
help
to
child
approaches
parent
get
Note : if help is sought and refused code rejecting
2) NEGATIVE
(NEG):
A) Harsh, punitive - typically involves aggressionor exaggerations of discipline to

involve killing, severe beatings.
Ex: - "I'm going to kick you"
Mother
throws
pot at child
-
(along
demeaning
blaming
"stupid
the
with
child
ex
girl")
- sexualizedaffection
B) Rejecting - parent pushes child away, rejects a bid for closenessor help.
Ex: - "That's an ugly picture"
Note if help is sought and refused code rejecting
.
Q Ineffectual - parent is unable or unwilling to help or assist the child when
the child ask a question or ask for help.
3) DISCIPLINEXONTROL
(D/C):
A) Discipline - involves a description of the parent as an authority figure who

disciplines and controls the child, child to parent, or may be parent to parent or child to
it
is
long
harsh.
May
cliild.
as
not overly
mvol--,'c physical punishment as
Fx: -I
told you NO! ", "Don't do that. "
311
APPENDIX
A. 2 (cont)
Ill. PERFORMANCE
A) CONTROL (CTR) - indicates attempts by the child to either control the situation
be
This
directive
through
the
may
examiner.
statements,attempts to distract the examiner.
or
Watch
for
the
the
out
examiner.
examiner saying "But in our story... "
or contradicting
Do not include changing the constraints of the story unless these are attempts to
See
Denial.
also:
manipulate.
Ex: - child says "No, his name is Harry"
to
child
continuing
reference a previous story while refusing to relinquish control
of a doll or prop
Code as follows after eachphaseof the narratives:

0= No controlling statementsor acts noted
One or more controlling acts or statements noted
Child must persevere to some extent.
B) EMOTIONAL OPENNESS - the following are over all ratings with the majority
being coded during the presentation, narrative development, and transition phases. Coding
schemeswill vary.
1) PERSONAL EXPRIESSIVITY - refers to verbal, non-verbal and physical
Hint:
judge
how
by
to
the child's expression is when
the
watch
responsesmade
child.
nothing much id happening.
EXPRESSION
OF
JOY
(JOY) - verbal and/or non-verbal acts on the part
a)
of the subject that depict some level of joy or pleasure. Score highest level observed at the
ciid of each narrative phase.
0= Neutral or bland expression, absenceof affect; or no positive is evident.
Maybe attentive to toys or person but no positive is evident.
I= Partial or fleeting smile - without eye involvement; or excitement in body
i-i-iovements
without accompanying warm smiles or positive vocalizations; straight mouth
include
positive interest or attentiveness (to people or toys) which
may
neutral
\,oc,-ilizations. Also include one instance of the child using a word depicting pleasure such
itglad" or "happy". Two or more words are coding as a 2).
"= Smile - Warm, full smile sustainedupturned comers for at least '/2second
body
by
and/or
pleasant
vocalizations
without
accompanied eye involvement
excitement or
Two
high
or more words indicating pleasure.
iioticeablv
arousal.
laughter
high
Vigorous
Laughter
smiles with
and/or other
pleasurable
3=
body
by
be
full
tension.
(ie..
pleasure
accompanied
screeches,
may
N-ocalizationsx-hich
Full
laughter).
tenuous
check and/or tiny chuckle.
or
312
APPENDIX
A. 2 (cont)
b) ANGER (ANG) - Score highest level observed

0= No Anger.
I= Slight/Some Anger - the subject's brow is drawn down at centre, mouth
is tight and taut; sustained briefly (2 -5 secs); may emit a staccato vocalization as in
is
by
instance
Slight/some
"bossy"
tone of
anger
portrayed
enacting
a
single
of
protest.
"
NO!
"),
(e.
doll
is
that
the
g.,
or
subject
staccato
voc
may
state
angry.
voice or single
2= Moderate Anger - the subject's display is the same as 1) but is sustained
longer and may include several staccato vocs; the subject may repeat brief displays
during
Moderate
is
by
the
times
two or three
coding
phase.
anger
several
portrayed
(e.
"Go
bed
No,
I
")
to
of
anger
vocalizations
g.,
now!
instances
won't!
3= Strong Anger - the subject's display is more intense than 2); may include
involve
is
differentiated
from
fullthan
this
staccato
rather
vocs
and
may
elongated
crying;
blow distress cry by: a) state preceding cry is anger than fear/grimace/whimper. Strong
by
having
instances
the
characters
repeated
angeris portrayed
of anger vocalizations that
into
If
be
spill
over
aggression.
aggression
may
is portrayed, strong anger should
scored
it
is
by
instances
if
two
three
or
preceded
of anger vocs.
only
highest
level
DISTRESS
Score
(DIS)
Do
knitted
brow
c)
observed.
score
not
here
by
This
below.
includes
category
unless
accompanied
affects
specific criteria
FEAR.
0= No personal distress or portrayed distress observed.
I= Slight distress or fear - child's eyes enlarge; brow raises, mouth opens
slightly; sustained for at least 2 seconds. Child may whine when describing source of
distress.Fear is portrayed by enacting a single instance of withdrawal of a character or
fear
don't
be
("I'm
"I
to
statementof
scared" or
alone! " or a "Whoah" as the character
want
is falling. )
2= Whimper - child's negative vocalizations are briefly sustained(2 -5 secs),
fke is showing extended grimace with brows drawn down and flattened; often seen as a
ti-aiisition before full- blown crying, which is never reached. This level should be scored for
portrayeddistress when charactersemit cry vocalization once or twice. Subject may simply
is
doll
that
the
verbalize
crying without the accompanying affect.
3= Full-blown cry - child's negativevocalizationsare extendedand intense;

browsdrawn down and flattened;mouth open;child's distresscommands response.This
level should also be scored for portrayed distress when character(s)emit several cry
becomes
distress
the
a salient theme,or when the subject
of a character
N,
ocalizatioiis or
I\Nveps.
c1le's
d) CONCERN (CRN) - Score highest level observedat the end of each
from
look
distinguished
be
by
it's
This
a
of
concentration
may
noting
iiarrative phase.
firimio,in relatioii to a concerningevent.This categoryis usedwhen the characteror subject
313
APPENDIX
A. 2 (cont)
is expressing or demonstrating concern for another person.

0= No concern - the subject displays pleasure, distress/fear, anger or
brow.
No
in
knitting
No
sympathy/concern portrayed
subject's narrative
neutral affects.
story.
I= Slight/some concern - the subject sobers (facial display changes from
down,
display
intense
focused
downward,
turned
to
slightly
mouth
gaze
or
another affective
for
knitted)
brow
(2
20
displays
knitted
time
an
extended
secs)
or
showing
not necessarily
brow expressionof moderate concern for a relatively brief time (I -5 secs). Subject portrays
instance
for
(having
doll
concern
of
a
character
a
express sympathy or
or verbalizes a single
fell
down.
include
"
"Uh
Also
"Aww,
oh"
she
child referencing the examiner
caring, e.g.,
during
the concerning portion of the story.
in a concerned manner
2= Moderateconcern- the subject sobersand incorporatesknitted brow

for an extendedperiod of time (5 - 20 secs);the child may display knitted brow for several
brief periods,alternating with neutralor otheraffectivedisplay.Moderateportrayedconcern
is seenincluding severalinstances remarking concernfor a character;an extendedtheme
is
in
discernible
the subject'sresponserather than a single remark.
of concern
3= Great concern - the subject displays the full sympathy face of mouth
for
3
in
drawn
down,
turned
a
minimum
of
seconds.
slightly
a
centre
and raised
eyebrows
If less,score as moderate concern. Great portrayed concern is seenin the subject's focusing
in
The
(perhaps
for
their
theme
the
narrative response.
only)
main
on concern a character as
involvement
display
include
extended
with the
several concern remarks and
subject must
caring/sympathy toward one of the characters.
SAD
(SAD)
the
and/or
child
exhibiting
of
sadness
any
vocalization
e)
in
drawn
brow
facial
their
and up.
a
expression with
Ex: - children stating they are sad becausemom and dad are leaving
Code as follows for each narrative phase:
0= No sadnessshown or mentioned.
I=
Uses a word connoting sadness (e.g., "sad" "not happy" "bad") or
displays sadness.
"He feels bad"
"lie's lonelv"
-1- More intense than 1. Prolonged expression of sadness(i. e., greater than
than
Uses
once.
more
sadness
connoting
word
a
seconds).
disappointment with a sigh
"He's very sad"
314
APPENDIX
A. 2 (cont)
f) ANXIETY BEHAVIOUR (ANX) - physical movements on the part of

indicates
dread.
includes
This
a
sense
of
apprehension
or
the child which
self-soothing
behaviours such as rocking or thumb sucking, fidgeting, chewing on either their lip or an
be
behaviours.
These
sporadic or continuous
may
object.
Ex: - subject chews on dolls head while examiner is talking
is
happen
to
next
- subject asking what going
hand
face
rubbing
body
motion
- repetitive
behaviour
self-soothing
hair
twisting
- pulling or
lip
sucking or chewing on
regressive or nervous speech
Code as follows for each narrative phase:
0= No anxiety behaviours noted
I= One or more anxiety behaviours noted
deny
by
ignoring
DENIAL
(DNL)
the
the
child may
conflict or story
g)
it or they may deny a part of the story that the examiner has presented. The subject must
demonstratean understanding of the conflict in order to be actively denying it.
Ex: - child tells examiner "He didn't hurt his knee" in the bicycle story
Code as follows for each of the narrative phases:
0= No denial, the child deals with the story as presented
I= Child initially resists dealing with the a horn or issue but then
(perhaps
it
after prompting)
addresses
Child denies a hom of the story throughout narrative
Ex. Mummy's headache goes away
Q ROLE OF PARENTS (RP) - Indicate which parent the subject utilized during
in
helping,
for
the action, or
the narrative phase
participating
acts such as consoling,
administrator or recipient of punishment or aggression.
Code as follovs under the over-all category once the transition phase is complete:
0= No parent present or at least one parent present but not utilized.
I= Mom is utilized, dad, if present, is not utilized. Code this when the examiner
has broucyht mom back into the story and the child character converses
Ldoll.
the
moni
Ith
315
APPENDIX
A. 2 (cont)
2= Dad is utilized,, mom if present, is not utilized.

Parents are utilized.
Code any of the above if the parents are just mentioned.

D) NARRATIVE
COHERENCE
(NC) - for four and five year olds
0= No responseor "I don't know what happens". Child may repeat a portion or all
of the story stem without any additions.
I=
Not coherent - fragmented shifts in story line. Child does not return to original
story stem.
2= Child stays within story line but does not addressthe conflict and/or story.
3= Child exhibits an understanding of the conflict but does not offer any resolution
is
when a resolution expected or does not offer an ending to the story. A portion
of the narrative may be incoherent.
Examiner might end up asking, 'So, what happened T
Ex., in Lost keys the keys are never found
Child handles the conflict by changing the constraints presented in the original
More
include
incoherent
Narrative
contradicting the story
stem or prompt.
shifts.
may
stem.
Ex, Mom says it is okay to have cookies.
The keys were never lost.
4=
5= Child demonstrates an understanding of the conflict or story and handles it

indirectly by offering an easier solution. More avoiding the conflict.
Ex., In lost keys child says, 'I lost the keys.'
In Crowd story child walks away without his/her ball.
6= Child demonstrates an understanding of the conflict or story and/or offers a
Typically
these stories are very short. Child
resolution without any story embellishment.
incoherent.
A
the
to
tell
the
the
story is
story. segment of
offers
minimal amount
7=
The same as 6 with NO incoherence.
Child demonstrates an understanding of the conflict or story and offers a

is
A
the
story incoherent.
segment of
t*esolLItIOII with some story embellishment.
The same as 8 with NO incoherence.
3316
APPENDIX
A. 2 (cont)
10 =A very coherent, logical, sequential series of events that are related to the story
but
does
Child
the
to
story
may add
not change the original story stem. An
stem.
understandingof the conflict and a resolution to the conflict are presented, or an
understandingof the story and an ending to the story are provided when there is no conflict.
There are no incoherent shifts in the story and there is a lot of embellishment.
E) DIRECT VS. INDIRECT PERFORMANCE STYLE (DI)

to
the
refers
degree of interpretation necessary on the part of the listener in order to understand the
being
For
direct
conveyed.
example,
message
a
perfon-nancestyle would be the child telling
the examiner "No, I'm not going to finish the story", while an indirect style would be the
down
looking
head
in
their
and
shaking
child
responseto the examiner's requests.This scale
between
discriminate
levels
inhibition
We
the
may
of
are trying to
across children.
determine the clarity of meaning here. Also, inhibition of action/play is what we are
trying to get at here but not of speech.
Ex: - child moves dolls but does not say anything
Code during the narrative development phase as follows:
0= No responseor there may be several things going on that are incoherent to
the examiner
I= Indirect (principally), when prompted child either provides no responseor
half a direct response (ie., nods head). You find yourself filling in the gaps of the story
2= Direct (principally), may be prompted by examiner. Very little if any
iiiterpretation is necessary on the part of the examiner. It is very clear what the child is
doing.
WITH EXAMINER - refers to the level of interaction the
This
include
the
throughout
examiner
all phases.
child initiates and/or maintains with
would
the child initiating a dialogue, answering questions, attending while examiner is talking,
involving
to
telling
the
the
the examiner
story examiner, or
VISMIIN,referencing
examiner,
frame.
This
the
in play.
narrative
may occur outside of
(RES) - this scale
1) CHILD'S RESPONSIVITY WITH EXAMINER
in
displays
degree
the
the
responseto the examiner's
assess
of enjoyment and eagerness child
is
Highest
7.
Intermediate
to
the
to
rating a
narrative story stems.
prompts
scale
respond
described
below.
If
be
all cues in the categor,
points should
used as well as anchor points
holistic
Utilize
in
back
then
this
a
approach when coding
up one point.
are not present
category.
Code as follows under the over-all category:
F) RELATIONSHIP
I= Child does not respond to examiner's prompts; no pleasure is seenand

is
distress)
from
(anger
frequent
outbursts,
subject's
gaze
averted
there are
or
away
negative
downward).
(eg..
gazing
examiner and materials
317
APPENDIX
A. 2 (cont)
display
does
but
.
negative affects (ie., mute refusal), score
If sub ect refuses to respond
not
2.
3= Subject reluctantly responds to coaxing by examiner; no pleasure is
distress)
infrequent
(anger
displays
toward
negative
or
affective
seen,primarily neutral or
is
directed
is
face.
to
to
and
materials
never
examiner's
the examiner; gaze averted
5= Subject respondsto examiner when prompted; occasionally smiles, but
directed
is
towards
the examiner; subject may
negative
affects
no
neutral;
primarily
affect
look to examiner's face once or twice during entire narrative; subject may hesitate but does
not refuse to respond.
7= Subject is ready to respond before the examiner invites responseand
displaysfrequent smiles and possibly occasional laughter; subject clearly enjoys the give
include
frequent
hesitates
to
take
respond; may
visual gazetoward examiner's
and never
and
face.
2) CHILD'S INVOLVEMENT OF EXAMINER (IE) - this scaleassessesthe
degreeto which the child initiates or sustains interaction with the examiner. The child may
in
her
to
the
assist or participate their narrative
examiner or may request
ask questions of
Consider
'neediness'
frame.
the
This
the
child's
narrative
may occur outside of
response.
level.
Code as follows in the overall column:
I= Subject does not invite or involve the examiner in any way during the
narrative.
3= Subject may invite or request assistanceor more information from the
during
the narrative; child may elaborate moderately; may offer toy or may
examiner once
look at the examiner several times. Subject may initiate a conversation.
Subject invites examiner to play a role or requests examiner's
in
during
the
response
narrative; child may elaborate extensively
assistance repeatedly
include
for
involvement.
May
the
also
to examiner's prompt and may end with a request
child initiating negotiations with the examiner.
IN PERFORMANCE (IP) - indicate under the Over-all
G) INVESTMENT
3
demonstrates
involvement
the
through-out
the
level
phases
subject
of personal
column the
h,recording the total number of investments noted. For example if the subject moves one
to
the
this
that
dolls
story
any
in
vvay
contributes
through
motions
more
or
one
or more
leave
has
If
in
to
investment
the
and
attempts
subject
performance.
one
would equal
demonstrated some investment in performance code a -I over the number of investments
(remaining
follows:
in
within
Examples
as
are
performance
of
investment
iloted, ie., -1/33.
the narrative)
5=
doll(s)
actions
other
the
no
performing
up
picking
just
(ie.
the
to
that
not
picking
storup
contributes
a
manner
in
props
- moving
318
APPENDIX
A. 2 (cont)
it
down
again)
a prop and putting
dolls
dramatisation
the
the
through
as
subject
such
moving
various motions
is
happening
telling
the
the
subject
examiner
what
narration,
for
in
different
dolls
the
talking
voices
effects
sound
adding
becoming
in
the
the
an
actor
subject
story
interjection
the
of
child's own related personal experiences
whispering
frame
the
narrative
- singing within
also
sound
effects
-
Count the number of investmentexamplesand code as follows in the over-all

column:
leave
Subject
do
(not
like
to
testing
to
=
attempts
area
something
purposeful
-1
bathroom).
the
to
go
0= No investment in performance noted, subject does not attempt to leave
1,2 3
Indicate
total
noted
of
number
investment
.
.....
1) INDICATORS OF THE CHILD UNDERSTANDING THE CONFLICT

(CUC) - Note that stories vary on whether they have one or two "homs" of a dilemma which
hom
is
"Presenting
dilemma,
homed
hom
For
the
one
of a2
each
is presentedto the child.
Problem" and one horn is the "Constraint. " The Presenting Problem is the quandary to
it
find
difficult
Constraint
The
find
is
to
is what makes
a solution.
which the child trying to
below,
Presenting
Problem
descriptions
Under
Problem.
the
Presenting
the
a solution to the
is indicated by (PP). Evidence is defined as verbalization or enactmentthat clearly indicates
dilemma.
horn
knows
the
the child
of
one
For 2 homed dilemmas:
0= No Evidence of either hom
Presenting
if
demonstrated
is
the
the
Evidence
I= Minimal
subjectunderstands
Problem only.
for
demonstrates
if
is
Evidence
the
Complex
evidence
2=
subject
scored
but
it
is
in
dilemma
the
horns
narrati--,
not
both
e.
the
point
some
at
of
Liriderstanding
into
first
bed,
Dick/Jane
NAP
(eg.,
and
gets
when
the
story,
in
iridicated simultaneously
toys).
the
toys,
with
the
plays
she
askedabout
For I homed dilemmas:
0= No Bidence of dilemma
319
APPENDIX
A. 2 (cont)
I= Complete Evidence is demonstratedif the subject demonstratesevidence

definition
One
is
Cooking
dilemma.
Story,
hom
to
this
the
the
for the one
exception
of
in
dilemmas
it:
is
burned,
2
(mostly)
has
the
embedded
child
orthogonal
and a mess
which
is made.The child's bum is considered the PresentingProblem,and if this is addressed,the
dilemma.
Evidence
For
Complete
the
the
of
representation
of
purposesof
shows
subject
if
it
is
Cooking
Story
homed
dilemma.
the
treat
as
a one
this scale,,we will
Descriptions of the one (or two) horns of the dilemma for each of the
Problem
in
dilemmas.
indicates
Presenting
(PP)
the
the
narratives.
SPILLED JUICE:
1. Child understands the juice needs to be cleaned up.

MOTHER'S HEADACHE
1. Child understands mother has a headache.

THREE'S A CROWD:
1. (PP). Child indicates that s/he understandsthat little sib wants to play.
2. Child indicates that s/he understands that peer refuses to have the
younger sib play.
BURNED HAND:
1. Child understands the burn needs attention. (Conflict does not

include the spilled gravy)
LOST KEYS:
finds
found,
be
lost,
keys
them.
to
Child
1.
or
or
need
are
understands
SWEET SHOP:
1. Child understands it was wrong to take chocolate/sweets.
SEPARATION:
1. Child acknowledges
them. Eg., Says or waves good-bye.
Eg.,
have
the
gone without
parents
accepts
or
REUNION:
have
that
the
Child
returned.
parents
1.
acknowledges or accepts
Says hello; greets parents in some way.
BATHROOM SHELF:
by
Child
sibling.
needed
1. (PP).
understands plaster is
2. Child understands prohibition against getting the plaster, or says the
plasters are on the shelf.
320
APPENDIX
A. 2 (cont)
EXCLUSION:
1. Child understands parents want to be alone and the child must
in
another room.
remain
BISCUIT TIN:
1. Child understands that the biscuits are forbidden.
321
APPENDIXA.
Child
No.
Rater
SJ
MH
3C
BH
LK
SS
SN
RN
BS
EX
BT
THEMES
_NO
COMPETITION
_SHARING
RIVALRY
_EXCLUSION
AGGRESSION
_EMPATHY
REFUSED
E/H
RESOLVE
_I-C
TATTLING
_COMPLIANCE
NON-COMPLY
ESCALATION
SHAME
AFFILIATION
BLAME
AFFECTION
TEASING
DISHONESTY
PUNISHMENT
REPARATION
VERBAL
[
CONF.
INJURY
ATYPICAL
REPETITION
REPRESENTATIONS
PARENTAL
SJ
IINEGATIVE
POSITIVE
DISCIPL&WIWG
MH
3C
I
BH
LK
SS
SN
RN
BS
EX
BT
322
APPENDIX
CTR
ID
PRE
STORY
NAR
A. 3 (CONT)
JOY
ANG
DIS
IP
cuc
CRN
SAD
ANX
CRN
SAD
ANX
CRN
SAD
ANX
CRN
SAD
ANX
t
RATER
RP
TRA
NC
DNL
ADR
I
D/ I
RES
J
CTR
ID
PRE
STORY
NAR
RATER
TRA
RP
NC
DNL
ADR
D/I
RES
CTR
ID
PRE
STORY
NAR
RATER
TRA
RP
NC
DNL
ADR
D/I
RES
CTR
ID
PRE
STORY
NARI
RATER
TRA
RP
NC
DNL
ADR
D/I
RES
IE
JOY
IE
JOY
IE
JOY
IE
ANG
DIS
IP
cuc
ANG
DIS
IP
cuc
ANG
DIS
IP
cuc
323
APPENDIX A. 3 (CONT)
CTR
ID
PRE
STORY
NAR,
RP
NC
DNL
ADR
D/I
PRE
STORY
NAR
RATER
TRA
NC
DNL
ADR
RES
IE
SAD
ANX
- T-
cuc
ANG
DIS
IE
IP
cuc
JOY
ANG
DIS
IP
cuc
ANG
DIS
RES
D/I
IP
JOY
CTR
ID
CRN
ID
CRN
SAD
ANX
CRN
SAD
ANX
CRN
SAD
ANX
PRE
STORY
NAR
RATER__
TRAI
RP
NC
DNL
ADR
D/I
ID
PRE
STORY
NAR
RATER
IE
RES
JOY
CTR
I
I
I
I
I
I
TRA
I
-NC
RP
DIS
CTR
RP
ANG
TRAI
RATER
I
JOY
DNL
ADR
D/ I
RES
IE
IP
cuc
II
II
324
APPENDIX
A. 3 (CONT)
CTR
ANG
DIS
IP
cuc
ID
PRE
STORY
NAR,
RATER
TRA.
RP
NC
DNL
ADR
D/I
RES
CTR
ID
PRE
STORY
NAR
RATER
TRA
RP
NC
DNL
ADR
D/I
RES
CTR
ID
PRE
STORY
NAR
RATER
TRA
RP
NC
DNL
ADR
D/I
RES
JOY
IE
JOY
IE
JOY
IE
ANG
DIS
IP
cuc
ANG
DIS
IP
cuc
CRN
SAD
ANX
CRN
SAD
ANX
CRN
SAD
ANX
325
APPENDIX A. 4
Frequency of Content and Performance Codes Observed
Table I: Frequency of the binary content codes no theme, competition, sharing, rivalry
(n
86).
in
the
story stems =
each of
and affiliation
Story Stem
Competition
Sharing
Rivalry
Affiliation
ap
ap
ap
ap
Spilled Juice
86
86
86
83
Mother's
Headache
86
86
86
83
Three's a Crowd
86
85
86
82
Burnt Hand
86
86
86
85
Lost Keys
85
86
86
83
Sweet Shop
86
85
86
86
Separation
84
86
86
84
Reunion
86
86
86
77
Bathroom Shelf
86
85
86
85
Exclusion
86
86
85
84
Biscuit Tin
86
86
86
86
a= absent
p= present
326
APPENDIX
A. 4 (CONT)
Table 2: Frequency of the binary content codes affection, compliance, non- compliance,
(n=86).
refused empathy/helping
Story Stem
affection
compliance
non-compliance
aP
refused
empathy/helping
aP
aP
Spilled Juice
82
85
85
86
Mother's
Headache
83
33
53
45
41
86
Three's a Crowd
82
81
86
86
Burnt Hand
81
84
80
83
Lost Keys
71
15
86
86
86
SweetShop
77
81
83
85
Separation
78
85
82
85
Reunion
65
21
86
86
86
Bathroom Shelf
74
12
58
28
32
54
85
Exclusion
73
13
58
28
28
58
86
Biscuit Tin
81
1
51
82
81
86
p= present
a= absent
Table : Frequency of the binary content codestattling, escalation of conflict, teasing and
dishonesty (n=86).
Dishonesty
Teasing
a
Escalation of
conflict
ap
Spilled Juice
85
85
86
84
Mother's
Headache
86
85
86
82
Three's a Crowd
85
83
86
83
Burnt Hand
86
86
86
81
Lost Keys
85
79
86
71
15
Sweet Shop
86
84
86
73
13
Separation
86
86
85
81
Reunion
85
86
86
84
Bathroom Shelf
78
86
86
71
15
xclusion
86
85
85
75
11
Biscuit Tin
57
29
1 86
85
63)
Story Stem
Tattling
327
APPENDIX
A. 4 (CONTJ
Table 3: Frequency of the binary content codes verbal conflict and personal injury (n=86).
Story Stem
Verbal conflict
ap
Personal injury
aP
Spilled Juice
86
84
Mother's Headache
86
86
Three's a Crowd
84
83
Burnt Hand
85
75
11
Lost Keys
71
15
85
SweetShop
83
81
Separation
86
82
Reunion
86
82
Bathroom Shelf
86
80
Exclusion
84
81
Biscuit Tin
85
1
11
84
a= absent
p= present
Table 4: Frequency of the complex content code exclusion in each story stem (n=86).
Story Stem
Exclusion
Exclusion of other
Exclusion of self
Exclusion of both
other and self
absent
Spilled Juice
66
16
Mother's
Headache
61
14
Three's a Crowd
36
20
29
Burnt Hand
71
12
Lost Keys
70
Sweet Shop
65
14
Separation
81
Reunion
Bathroom Shelf
76
Exclusion
80
Biscuit Tin
60
2)
328
APPENDIX
A. 4 (C
Table 5: Frequency of complex content code aggression (n=86).

Story Stem
Aggression
Physical
Verbal
absent
aggression
aggression
Both physical and

verbal aggression
Spilled Juice
79
Mother's
Headache
80
Three's a Crowd
62
19
Burnt Hand
80
Lost Keys
62
21
Sweet Shop
72
12
Separation
77
Reunion
74
10
Bathroom Shelf
81
Exclusion
74
Biscuit Tin
76
Table 6: Frequency of complex content code empathy and helping (n=86).

Story Stem
Absence of
helping/empathy
Empathy
Helping
Both helping and

empathy
Spilled Juice
83
Mother's
Headache
46
40
Three's a Crowd
58
28
Burnt Hand
16
69
Lost Keys
82
Sweet Shop
Separation
82
Reunion
Bathroom Shelf
11
74
Exclusion
85
Biscuit Tin
85
10
10
329
APPENDIX
A. 4 (CONT)
Table 7: Frequency of the complex content code interpersonal conflict resolution (n=86).
Absence of interpersonal
conflict resolution
Seeks help
Spilled Juice
86
Mother's
Headache
77
Three's a Crowd
57
20
Burnt Hand
86
Lost Keys
84
Sweet Shop
86
Separation
85
Reunion
86
Bathroom Shelf
82
Exclusion
86
Biscuit Tin
86
Story stem
Uses adult
Uses threats
strategy
Table 8: Frequency of the complex content code shame (n=86).

Story Stem
Absence of shame
Shames self
Shames other
Shames both self

and other
Spilled Juice
71
14
Mother's
Headache
82
Three's a Crowd
78
Burnt Hand
67
19
Lost Keys
78
Sweet Shop
65
20
Separation
84
Reunion
85
Bathroom Shelf
72
13
Exclusion
79
Biscuit Tin
73)
12
330
APPENDIX
A. 4 (CONT)
Table 9: Frequency of the complex content code blame (n=86).

Story Stem
Absence of blame
Blames self
Blames other
Blames self and

other
Spilled Juice
86
Mother's
Headache
85
Three's a Crowd
86
Burnt Hand
80
Lost Keys
73
Sweet Shop
83
Separation
86
Reunion
86
Bathroom Shelf
84
Exclusion
86
1 11
Biscuit Tin
1 83
Table 10: Frequency of the complex content code punishment (n=86).

Absence of
punishment
Verbal
Physical
punishment
punishment
Both verbal and

physical punishment
Spilled Juice
43
23
12
Mother's
Headache
63
16
Three's a Crowd
77
Burnt Hand
43
40
Lost Keys
76
Sweet Shop
28
41
10
Separation
79
Reunion
79
Bathroom Shelf
4
-5
22
Exclusion
70
14
Biscuit Tin
37
333
Story Stem
7
0
18
18
331
A. 4 (CONT)
APPENDIX
Table 11: Frequency of the complex content code reparation and guilt (n=86).
Story stem
Absence of reparation
and guilt
Reparation
Guilt
Both reparation
and guilt
Spilled Juice
10
63
10
Mother's
Headache
86
Three's a Crowd
83
Burnt Hand
71
11
Lost Keys
33
46
Sweet Shop
44
38
Separation
85
Reunion
86
Bathroom Shelf
81
Exclusion
84
Biscuit Tin
75
51
61
Table 12: Frequency of the complex content code atypical response (n=86).
Story Stem
Absence of atypical
response
Neutral or
positive
Negative
Both neutral or positive

and negative
Spilled Juice
75
Mother's
Headache
84
Three's a Crowd
82
Burnt Hand
73
Lost Keys
70
12
Sweet Shop
73
Separation
76
Reunion
76
Bathroom Shelf
78
Exclusion
74
Biscuit Tin
78
11
31
A. 4 (CONT)
APPENDIX
Table 13: Frequency of the complex content code repetition (n=86).

Story Stem
Absence of
repetition
Repetition of
same story
Repetition of
prior story
Repetition of both same

and prior story
Spilled Juice
78
Mother's
Headache
83
Three's a Crowd
85
Burnt Hand
79
Lost Keys
82
Sweet Shop
83
Separation
81
Reunion
86
Bathroom Shelf
80
Exclusion
83
82
1
2
1
Biscuit Tin
Table 14: Frequency of the complex code positive parental representationss (n=86).
Absence of positive
parental
representation
Positive mother
only
only
Both mother and

father
Spilled Juice
82
Mother's
Headache
83
Three's a Crowd
75
39
31
13
75
Sweet Shop
80
Separation
85
19
55
330
Exclusion
80
Biscuit Tin
78
Story Stem
Burnt Hand
Lost Keys
Reunion
Bathroom Shelf
Positive father
64
ii
A. 4 (CONT)
APPENDIX
Table 15: Frequency of the complex code negative parental representations (n=86).
Story Stem
Absence of negative
parental
representation
Negative mother
only
Negative father
only
Both mother and

father
Spilled Juice
79
Mother's
Headache
83
Three's a Crowd
84
Burnt Hand
76
Lost Keys
77
SweetShop
75
10
Separation
85
Reunion
83
Bathroom Shelf
85
Exc usion
83
Biscuit Tin
83
2
-0
Table 16: Frequency of the complex code disciplining parental representations (n=86).
mother only
Disciplining
father only
Both mother and

father
44
12
21
Mother's
Headache
60
26
Three's a Crowd
72
Burnt Hand
47
1
-3
Lost Keys
75
Sweet Shop
29
56
Separation
86
Reunion
79
54
67
1 13
19
1
Absence of disciplining
parental representation
Disciplining
Spilled Juice
Story Stem
Bathroom Shelf
Exclusion
Biscuit Tin
22
334
APPENDIX
A. 4 (CONT)
Table 17: Frequency of role of parent in each story stem (n=89).

Story Stem
no parent
indicated
mother only
indicated
father only
indicated
both parents
indicated
Spilled Juice
13
11
58
Mother's Headache
78
41
12
32
Burnt Hand
42
39
Lost Keys
78
SweetShop
74
10
Separation
56
32
Reunion
10
75
78
Exclusion
13
12
61
Biscuit Tin
91
1
23
Three's a Crowd
Bathroom Shelf
61
1
51
Table 18: Frequency of child's understanding of the conflict in each story stem (n=89).
Story Stem
no understanding
some understanding
Spilled Juice
10
79
Mother's Headache
10
79
Three's a Crowd
25
64
83
15
74
SweetShop
84
Separation
86
Reunion
81
Bathroom Shelf
11
78
Exclusion
24
65
Biscuit Tin
83
Burnt Hand
Lost Keys
335
APPENDIX
A. 4 (CONT)
Table 19: Frequency of indirect vs direct performance style in each story stem (n=89).
Story Stem
indirect style
direct style
Spilled Juice
14
75
Mother's Headache
81
Three's a Crowd
81
Burnt Hand
80
10
79
SweetShop
80
Separation
80
Reunion
81
Bathroom Shelf
81
Exclusion
12
77
Biscuit Tin
9
1
80
Lost Keys
Table 20: Frequency of types of response to examiner in each story stem (n=89).
Story Stem
no response
reluctant response
ready or enthusiastic
response
Spilled Juice
21
64
Mother's Headache
17
69
Three's a Crowd
14
72
Burnt Hand
19
67
Lost Keys
18
68
Sweet Shop
25
61
Separation
22
64
Reunion
19
66
Bathroom Shelf
15
71
Exclusion
Biscuit Tin
1 19
65
66
1
336
APPENDIX A. 4 (CONI)
Table 2 1: Frequency of level of involvement of examiner in each story stem (n=89).
Story Stem
little or no involvement
moderate or more involvement
Spilled Juice
47
42
Mother's Headache
42
47
Three's a Crowd
46
43
Burnt Hand
41
48
Lost Keys
52
37
Sweet Shop
52
37
Separation
47
42
Reunion
60
29
Bathroom Shelf
44
45
Exclusion
47
42
Biscuit Tin
68
21
Table 22: Frequency of number of investments in performance in each story stem (n=89).
Story Stem
three or fewer
four
five
Spilled Juice
31
33
17
Mother's Headache
24
38
22
Three's a Crowd
24
23
Burnt Hand
28
23
27
11
Lost Keys
34
32
16
Sweet Shop
20
25
35
Separation
Y)
30
27
10
Reunion
21
35
26
Bathroom Shelf
15
30
32
12
Exclusion
16
Biscuit Tin
28
-38
six or seven
39
23
337
APPENDIX
A. 4 (CONT)
Table 23: Frequency of denial in each story stem (n=89).

Story Stem
no denial
some denial
Spilled Juice
59
30
Mother's Headache
64
25
Three's a Crowd
56
33
Burnt Hand
51
38
Lost Keys
47
42
SweetShop
73
16
Separation
77
12
Reunion
77
12
Bathroom Shelf
68
21
Exclusion
54
35
Biscuit Tin
78
1
11
1
Table 24: Frequency of ada-ptivenessof response in each story stem (n=89).

Story Stem
low level of
adaptiveness
moderate level of
adaptiveness
high level of
adaptiveness
Spilled Juice
26
53
10
Mother's Headache
29
46
14
Three's a Crowd
33
49
Burnt Hand
36
41
12
Lost Keys
50
36
Sweet Shop
39
43
Separation
33
46
10
Reunion
41
31
17
Bathroom Shelf
23
50
16
28
Exclusion
Biscuit Tin
36
145
18
338
APPENDIX
A. 4 (CONT)
Table 25: Freciuencv of narrative coherence in each storv stem (n=89).

Story Stem
conflict not
addressed
Spilled Juice
16
Mother's Headache
conflict partially
addressed
conflict addressed
some resolution
conflict addressed
and resolved
52
21
12
23
32
22
Three's a Crowd
10
37
20
Burnt Hand
17
36
32
Lost Keys
28
25
19
17
SweetShop
17
11
38
23
Separation
21
23
45
Reunion
23
45
18
Bathroom Shelf
14
13
25
37
Exclusion
39
18
18
14
Biscuit Tin
17
10
39
23
Tqhh- ? Cv Frenuenrv
(n=89).
in
iov
storv
stem
each
each nhase
and anger
of controlControl
Story Stem
p
Spilled Juice
Mother's Headache
Three's a Crowd
Burnt Hand
Lost Keys
Sweet Shop
Separation
Reunion
Bathroom Shelf
Exclusion
Biscuit Tin
Joy
n
Anger
t
50
66
26
20
15
56
68
24
14
46
66
21
14
13
12
12
51
63
27
11
14
11
42
71
26
18
10
47
62
28
27
17
46
67
11
11
31
62
28
10
12
40
66
36
20
14
50
66
33
41
1
65
19
16
16
8
12
14
7
1
7
phase
narrative
n=
phase
p= presentation
t= transition phase
3339
APPENDIX A. 4 (CONT)
Table 27: Frequency of distress, concern, sadnessand anxiety in each phaseof the story
(n=89).
stem
Distress
Story Stem
p
Concern
t
pn
Sadness
Anxiety
pn
Spilled Juice
44
69
4-5
Mother's Headache
63
67
44
Three's a Crowd
12
10
55
66
45
Burnt Hand
20
61
62
50
Lost Keys
10
59
68
46
Sweet Shop
57
60
49
Separation
60
58
Reunion
33
52
38
Bathroom Shelf
29
62
63
50
Exclusion
66
64
49
Biscuit Tin
12
p= presentation phase n= narrative phase
55
t= transition phase
54
1
223
340
APPENDIX A5
Repeatedmeasurest-test
to seeif the meansbetween
the stories for the samefactor
different from each
other
are
Table T-values for Factor I
(storytelling/relationship
to task) betweenthe II
(n=89).
H SJ
F1 MH
F1 3C
F1 BH
F1 LK
F1 SS
F1 SN
F1 RN
stories
H BS
FI EX
F1 SJ
H MH
5.49***
F1 3C
1.60
F) BH
1.07
F1 LK
F) EX
3.81***
-0,18
-1.09
-0-65
0.18
3.20*
4.38***
5.87***
-0.43
2.26
2.86**
3.76***
-4.54***
0.49
-4.90***
3.57**
0.71
3.23*
01
001
Table
-0.43
4.97***
5.48***
F1 SN
F1 BS
-3.80***
0.52
F) SS
F) R
-3.43**
-0.66
-1.41
-1.24
-0.73
-5.30***
4.75***
-1.00
-0.76
-0.06
-5.64***
-5.02
-2.65*
-3.17**
0.95
0.34
-0.68
3.96***
-1.86
0.35
-0-06
3.37**
<. 000
0.82
-4.50***
4.30***
-1.54
-0.92
-2-96*
T-values for Factor 2 (discipline

and punishment) of II stories (n-86).
F2 SJ
F2 MH
F2
3C
F2 BH
F2 LK
F2 SS
F2 SN
F2 RN
F2 BS
F2
F2 SJ
F2 MH
4.71
1:2 3C
4.82***
F2 BH
1.32
0.63
-3.20**
-3.46**
F2 LK
S.54***
1.37
0.89
4.
-Sl***
F2 SS
57
-.
-5.76***
-6.37***
-2.36*
-6.48***
8.00***
473***
4.16***
7.32***
3.12
10.20***
3.50**
2,93 *
6.20***
1,84
9.08***
80*
--)
3.37**
-0.37
75***
-6
F2 SN
F2 RN
1`2BS
F2 EX
F2 BT
7 40***
2.79**
4.86***
94
-.
-1.69
1.01
-5.78***
7*
-2.3
1 40
0.42
4.00***
47***
-6
2 46*
-1.52
-6.04
4*;
***
-4
6.54***
26*
-3
-0 40
-10.12***
12
2.46 *
-2
111111111111
..................
-3.93***
-6.830
341
000
**<. 001
T-values for Factor 3 (negative/controlling) between the II stories (n=89).
Table
F3 SJ
F3 MH
F3 3C
F3 BH
-0.89
0.06
0.98
-2.08
1.55
-
F3 LK
F3 SS
F3
SN
F3 RN
F3 BS
F3
UX
F3 SJ
F3 MH
1-0.10
F3 3C
-1.46
-1.86
F3 BH
-2.89*
-2.60*
F3 LK
-1.42
F3 SS
-1.38
-3.55**
SN
_F3
F3 RN
-3.76***
-2.98*
-3.05*
-2.75*
F3 BS
-3.18*
-1.35
-1.30
-0.56
-0.52
-2.49*
-1.41
0.89
-1.46
0.90
0.04
-0.41
-0.44
-0.10
-0.08
-3.07*
-3.38**
-1.79
-0.86
-2.02
0.57
F3 EX
-2.95*
3.16*
1.33
0.44
1.26
[E3 BT
-0.98
1.97
-
2.44
-0.36
0.48
'Fable
10.44
-1.86
-1.08
-1.13
-0.50
-1.42
0.95
**<. 001 *** <.000
T-values for factor 4 (positive/affection) between the II stories (n=86).

F4 SJ
F4 MH
00
.
65
-1 .
-1.73
F4 3C
F4 BH
F4 LK
F4 SS
F4 SN
F4 RN
1.4 BS
F4 FA
F4 SJ
F4 MH
F4 3C
F4 BH
F4 LK
F4SS
-5.77***
*
-3.03
-2.18
F4 SN
-0.65
F4 RN
*
-4.74
F4 BS
-5.60***
F4 EX
F4 BT
*,
01
-1.89
1
-1.09
**<. 00 I
-5.96***
-3.40**
-1.58
-0.69
1
-4.3
-5.60***
1
2.35
11.04
***
<. 000
-4.05***
-1.83
00
.
1.04
-4.16
4.49***
0.49
.1
-0.55
1.86
3.94***
1.74
4.74***
2.55*
-0.34
-2.09
-0.47
1
-3.63***
4.25***
-2.35
1
-1.21
1
*1
-3.13
0.93
-3.46**
4.60***
1
0.48
-0.62
-4.53***
-5.06***
1
1.58
1
0.39
-0.12
-2.78*.
-3.80***
1
-3.23*
4.41 ***
1
0.97
342
APPENDIX
.1
Mill Hill Vocabulary Scale
A SHORT VOCABULARY SCALE

for Clinical Use with Individual Patients
Prepared by J.C. Raven
(Each word correctly explained is equivalent to a score of 5 on the whole. M. H. V. Scale)
Name
Age
.................................................
..................
Sex
Date
........................................................................................................
.........................
Other Assessments
...................................
Cruel
Near
Shrivel
*a
Ws
C
0
View
C
5
V
Liberty
M
cc
460
10-=-Mingle
Elevate
25 --f-Verify
P.T
343
APPENDIX B-lJCONT)
(7-S
-S
I-
No
344
APPF. NI)IY
R, 7
374
DIFA"T nVESTIOTIAIRE
Part I.
Your name
Address
Phone No
Today s date
**
Part
1.
11.
On the following
the number that is xost typical
questions
please circle
1#jJW:
k
of your baby.
h(p4
thIn.
the typical-", biby.
you
ut averate"
means
be scared.
vould
How easy
upset?
or
difficult
Is
it
for
.12345
very easy
2.
** ** **
you to calm
soothe
or
is
it
for
1234567
very ezz
M
baby
u,,
when your
you to predict
about average
'F.aw easT or diff,

hunp.--7
'cult
is
V-'Ien your babY vin'i
-Far you to predict
4.
easv
about averalre
cf-sy or diff!
cU"
ha/she c-. jas or fusses?
is
it
for
r4
.0 lb
easy.
bothari==
1-2 ti=ez
per day
E,0169mmich docs your
3-4 ti=as
per day
tizes
-5-6
per day
the
difficult
...
.ca
U,:
difficult
'
etc. ;
irritablcand
6
10-14 ti--ei'limri.
per day
-'tha=
baby cry and fuss in ; enersl?

6
23
Ver: -I li: tle:
MUC11 less man
averzrc b,bv
.5
7-9 tizes
per day,
go to sleep
hc:;
b-OF
:
Your
fussy
baby
r-Ov
does
get
your
mas per day, on the average,
for either-short-or
long pq#ods of time?
123
r. "er
6.
L-mcr-VI.
-.ztls
tL'
YOU,
4
.5
about average.
23
very
Is
7
dif if Icult
23
,Y
baby when he/she
about average
Hcm easy or difficult

and wake ujo?
ver-
your.
averaga Amount:
about ar ouch
Air the averAxe
b 2,----
7
lot;
much
s
wre than
tho avcra; c
baby
345
APPENDIX B.2CQNT
'Raw did your baby respond to his/her

123
very wellbaby loved

S.
baby'respond
to his/her
123
ire%T favorablyliked It Inzediately
9.
Eow does your baby typically
solid
food?
6
7
very nexativ! elydid not like At
at all
-n st always
respornds favorably
Uaw wall
does your
123
very well.,
always likes
eventually
respond to being in
baby adapt to things
it
infant
Eow does
1
vary walllikas
it
bab7
2
react
7
alln-st always
responds nagatively
at first
7-10) evc==AUj?
6
7
almost always
it
dislIkas
In the end
7
very &as ily
upset, by
'
t1himus that
It
bo tber
wouldn
t babies
get upset?
during
when you ara

4
drassin;
about averarmdoesn't
mind it
dimeringe
6
45
moderate intensit7 or loudness
23
Intensity
your
new place?
(such as in ita=
(e. g., before fa*ding,

When yoar baby gets'upset
fuss?
halsha.
does
loudly
and
cry
vigorous17
or
very mild
or loudness
P_
45
about average
123
very hard to
by
upsat-aven
thium
thet upset
most babies
7
AfEwst always
neg&respand
tively
at first
45
ends uv 'Ming
&
it aborut half
the ti=e
Aoes your
easily
45
responds favorably
about half the tiza*
A'
14.
teiriblydidn't
likelc
45
responds favor bly
about half the t:i=e
23
13.
f irst
Flow does yorur baby typica. 1.1.1y

respond to a new person?
at always
respornds favorably
11.
45
neither liked
it
nor disliked
23
10.
bath?
45
liked
neither
it
nor disliked
it
How did your
first
)
-how
etC.
7
very lafid or intansa., -really
cut laose
hi=/her?
5
doeim't
at all
Uke
it
346
APPENDIX
Raw active
is
your
baby in general.?
12
vez7 calm and
quiet
16.
B. 2 CONI)
345
67
average
very active
and vigorous
How much does your baby smile and 1m,

" ke happy sounds 7
34
.5
an average WDount
a great deal.
zm=h more than
it Infants
17.
What kind
of mood is ycur baby generally

2
Wery littlet
Much less than
Most 4nfAMtS
in!
345
very happy and

cheerful
18.
7
serious
neither serious
nor cheerful
How much does your baby enj cy playinr,

2
little
games with
you?
6
345
a great deal,
loves it
really
7
vinT littlep
doesn't like
about average
it
19.
How much does your
baby want
20.
21.
30CetJ--PS
held;
B=
easy
is
it
for
vary
about
you to predict
used a diaper
3.4
changs?
7
very difficult
is your baby's mood?

6
345
about invelege
cl, v,n as seldom.

=d changes slowly,
vhct= he/she does rh pnze
bec=e when peo; le play -,elth
does your bab

3
'?a"--7 e=citad
baby will
about average
Vy excitad
-ow
7
vorably
-very -0
gets quite upsat
average
vhe= 7our
about tver==m
S
,
everyday-routinsi,
345
G&SY
Bow ch=eable
7
a great deal--wmt
to be hold el=st
all the tim
w2nts to be
somet: i=es not
How does your baby respond to disruptions

and changes in the
such as when you go to church or a meeting,
on tSips . etc. ?
7.1.
45
12
very favorably
doesn't
get upset
mnCh
to be held?
2
wants to be free
=ost of the ti=e
VO.Ly
7
chan=as oft2n
=d rupidly
or talk
to him/her?
7
not at
347
APPENDIX B. 2 ( ON]: )
rate the crwerall

24.6jume
average mother.
degree
of difficulty
your
baby vould
present
fcw the
234567
super
easy
ordInarys
-problems
some
Ugbly. difficult to deal

with
348
APPENDIX B.3
The Bus Story Test
Ile
Bus Story
Page I
Once upon a time there was a very naughty bus.
Wbile his driver was trying to mend him, the bus decided to run away.
lb.
Aj
41vz
0* --.
Page 2
He ran along the road besidea train.

faces
fiuiny
other.
each
raced
mid
other
each
at
made
lley
into
because
train
the
a tunnel.
But the bus had to go on alone,
went
Ile hw-riedinto
"Stop,
his
blew
shouted,
and
lie
whistle
who
poficenian
the city where met a
bus".
,.
-- .-Z-, 7 10
.---..
.
_c
-.
%.
"
'"
-.
/
1, -,
--
eb-
le
349
APPFNDTY R UC-QNT)
Page 3
But the naughty-bus paid no attention and ran on into the country.
fie said, "I'm tired of going on the road".
So lie jumped over a fence.
He met a cow who said, "Moo, I can't believe my eyes".
A. & v
iP*1
%,.,
-,
': 0. ,"*0..
""
e;
0",,
.01
Page
The bus raced dowti the hill.

As soon as he saw there was water at the bottorn, he tried to stop.
But he didii't know how to put on his brakes.

So lie fell in the pond with a splash and stuck in the mud.
Wheu the driver fowid where lie yvas,lie teleplioned for a crane to pull him out and put bim
back on the road again.
..........
--
-------------
'if
=: 4
=-
III
"May
Y'
If
==
-II
fI
350
APPENDIX C
Child BehaViour Checklist
For
describe
items
item
that
that describes your child now or within the Past 6 months.
children.
list
each
of
is a
pie
Is
Circle
if
is
Item
true
Item
1
true
the
the
very
or often
of your child.
somewhat or sometimes true
the 2 If the
of
V
true
Please
items
Is
0.
the
Item
of
your
child.
circle
If
as
not
as
all
well
you can. even some do
answer
11the
not st
jV, to your child.
O= Not True (as far as you know)

2
2
1.
2.
I =Somewhat or Sometimes True

10
Acts too young for his/her age

Allergy (describe):
Argues a lot
Asthma
2 =Very True or
Often
31.
Fears he/she mioht think or do

sOfnethin(
bad
)
2
?,
,0
0
!1- ) 2
,
32.
33.
Feels he'she fts to be perfect

Feels cw complains that no one loves hir
12
0*
12
34.
35.
Feels others are out to get hiry

hirnfher
Feels worthless or
infericw
o( inferior
12
36.
37.
Gets hurt a lot. accident-prone

Gets in many lights
12
2
2
5.
6.
Behaves like opposite sex

Bowel movernents outside toilet
2
2
7.
8.
012
Bragging. boasting
Can't concentrate. can*t pay attention (of long
12
12
38.
39.
Gets leased a lot

Hangs around with Children
children who
wt get in
trouble
t(ouble
9.
Can't get hlsIher mind

(descrit>e):
obsessions
off certain
12
40.
Hears sounds or voices that a

afen*t lhc(describey
Can*t sit still.
or hyperactive
10.
restless.
012
thoughts.
DO
0
11.
12.
10
13.
14.
to adults or too dependent

Complains
of loneliness
Confused or seems
Cries a lot
Cruel to. animals

Cruelty. bullying.
2
2
17.
18.
Day-dreams or gets lost in his/her thoughts I

Deliberately harms self or attempts suicide
2
2
19.
20.
Demands a lot of attention

Destroys his/her own things
21.
Destroys things
23.
Disobedient
24.
Doesn't
25,
Doesn't
26.
Doesn't
belonging
45.
Nervous.
highstrung.
or tense
46.
Nervous
movements
Cx twitching
D12
C90-
47.
Nightmares
012
48.
0)
to his/her family
eat well
28,
Eats or dirinks things

don't include
sweets
Fears
other
00
situations.
animals.
(describe)
school
ceflain
than
12
49.
12
so.
Too (earful
12
51.
Feels dizzy
12
52.
12
53.
Feels too guilty

Overeating
12
54.
55,
Overilred
Overweight
56.
Physical
CO
that are not food (describe):
or places.
1
1
)
0,/ 1
2
2
2
1
1
1
2
2
2
0)
.0
0
(des(
Not liked by other children

Constipated.
doesn't move bowels
12
get along with other children

seem to feel guilty after misbehaving
Easily jealous
29.
at school
27.
Bites fingernails
or meanness to othefs
Likes to be alone
Lying or cheating
44.
or other children
Disobedient at home
42.
4 3.
2
2
iS.
16.
22.
ImpulSive or
thinking
Impulsive
aCtS without
(X acts
without think
to be in a tog
2
2
1'.
Clings
0 Qj
2
2.4
or anxious
p(oblems
wilhout
known
causea. Aches or pains

Headaches
b.
C. Nausea. feels sick
Problems with eyes (desc(lbe)
d,
e.
f.
0.
Rashes or other skin problems

Stomachaches or cramPS
vomIIinQ. throwing up
351
APPENDIX C (CONT-)
57.
2
58.
Ph
aR47*#IlaCKS
Cicksol
_)tALC:
skin.
no4-
412
peopie
or
other
Parts
of
84.
Strange behavoor (describe):
85.
Strange
86.
Stubborn.
87.
88.
Sudden changes
Sulks a lot
-
89.
90.
Suspicious
Swearing or obscene language
2
2
91.
92.
Talks about killing self

Talks or walks In Sleep (describe):
2
2
93.
94.
Talks too much

Teases alo(
2
rw
'g.. 1 2
-.
95.
96.
Temper tantrums or hot temper

Thinks about sex too much
97.
98.
Threatens people
Thumb-sucking
2
12M.
99.
Too concerned with heatness or clea(

Trouble sleeping (d escdt>e)-
2
2
101.
102.
Truancy. skips school

Underactive. Slow Moving. or lacks er
2
2
103.
104.
Unhappy. sad. or depressed

Unusually loud
105.
Uses alcohol of dru' for nonmedic.

purposes (descrit>e):
W6.
Vandalism
107.
108.
Wets self du(ing

Wets the bed
109.
Whining
12
110.
Wishes
12
Ill.
Withdrawn.
12
112.
Worrying
113.
Please
body
(describe):
P2
12
I2
12
12
12
2
2
/1
%1
12
Plays with own sex parts in public

Plays with own sex pans too much
61.
Poor school
62.
Poorly coordinated
63.
Prefers
playing
64.
Prefers
playing
65.
Refuses
66.
Repeals
67.
G8.
69.
70.
12
12
12
2
!
i
59.
60.
work
012
-)q 12
or clumsy
with older children

with younger children
certain acts over and over

(describey
compulsions
Runs away from home

Screams a lot
Secretive. keeps things to self
Sees things that aren*t there (decr; be):
Self-conscious
72.
Sets fires
74.
to talk
71.
73.
D12
o
01
0
01200
off 0( Clowning
12
12
75.
76.
Shy or timid
Sleeps less than most children
77.
Sleeps
0)
more than most children

and/or night (describe):
during da)
12
912
110
12
002
02
1
1
78,
Smears
Or plays
wilh
79.
Speech
pfoblem
(OescobeY
80.
Sta(es
bo--ovel movements
(9
blankly
that
81.
Steals
82.
Steals
83.
Sjo(CS
at home
outside the home
up
Or Irritable
sullen.
In mood of' feelings
or eady. embarrassed
Sexual p(oblems (desc(ibt-):
Showing
ideas (describe):
M4os
h(! /she
oi2
d(>,esn*l
nef
012
the day
to bc ol opposite
doesn*t
w(ite
we(e
not
sex
get involved
in any
problerns
listed
at>ove:
with
YOuf ct
352
APPENDIX
il I
AAI Questions
55
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Oriented re family, where you lived, moved much, what family did for living?
-Grandparents seen much, or died when parents young--know much about grandparent
who
died before your birth? --- Other persons in household?
Sibs
(Keep
nearby?
--short/demographic, no more than 2 or 3 minutes).
I'd like you to try to describe your relationship with your parents as a young child if
you
...
far
back
start
as
as you remember?
could
Five adjectives describing relationship with mother, as early as you can remembmr
but about
5-12 is fine (write down adjectives). Probe each in sequence given, asking for memories,
incidents for each. When one specific incident is
briefly
given,
seek a second. When
is
used for a first adjective, repeat query with reference to original
another adjective
When
general or scripted memories are given, give a final probe for a more
adjective.
Whenever a specific incident is given, however, enquire briefly for a
specific memory.
incident.
second
Five adjectives father. As above.
To which parent closest, and why? Why not same feeling with other parent?
When upset as a child, what do? Pause. (a) Upset emotionally? --incidents? (b) Physically
hurt-- incidents? (c) When ill--what would happen?
First separation? How did you respond? How did parents respond? Other separations that
stand out?
Felt rejected as a child? How old? What did you do? Why parent did these things? Realize
he/she was rejecting you?
Were you ever frightened or worried as a child?
8a.
Parents ever threatening--for discipline, jokingly? (Elective per researcher: Select one
happened
).
in
to
form
you?
of punishment used researcher's community--ever
specific
Some people have memories of some kind of abuse in family--happen to you or in your
family? --- what exactly happened, describe--how old, how severe, how frequent? ---this
to
child?
approach
adult?
as
you
experience affect
--affect
In general, how do you think your overall experiences have affected your adult personality?
development?
to
Any aspects of early experiences you consider a set-back
your
Why do you think your parents behaved as they did, during childhood?
important
though
Or
not
like
especially
Other adults close
aults
other
parents as a child?
household?
in
live
responsibilities?
caregiving
they
(Your
time--did
---had
at
age
parental?
).
important?
why
Loss of parent, other close loved one (sibs) as child? --- circumstances? ---age? --how
feelings
felt
how
time?
over
changed
at
recall
--respond at time? --- sudden or expected? --lost,
(If
like?
it
parent
on
remaining
effect
sib
or
funeral?
parent
was
time? --- attend
--- what
to
loss
child?
own
on
approach
personality?
this
)
adult
household?
on
--of
effect
and on
--13a.
Other losses in childhood. Queries as above
13b.
Important losses in adulthood. Queries as above.
traumatic?
potentially
as
--after participant
Ever had any other experiences you regard
and
overwhelmingly
rare
herself,
mean
you
clear
interprets for himself or
make
judgment.
best
immediately terrifying events--probe using
Elective per researcher.
between
and
childhood
parents
in
with
Were there many changes your relationship
adulthood?
with
contact
for
much
an
adult?
as
like
you currently
What is relationship with parents
dissatisfaction?
of
like
sources
current
is
currently?
relationship
parents at present? what
of satisfaction?
Id
time
sufficient
imaginary
after
(or
child)
oI
year
one
from
--Feel now when separate
child?
(imagined)
feel
Do
about
worried
ever
you
describe
add
response,
has passed for subject to
child?
like
future
to
kind
see
you'd
Thinking
of
of
from
now, what?
If 3 wishes for child 20 years
ink.
th
minu
tw,
to
te
fby-Ctittd-ur
,
here
I'm
you
of
something
thinking
frdp
learned
own childhood experience?
Any one thing
had
L-ind
vou
thp
nf
rhildhnnd
+
--
APPENDIX D. 2
Means and SDS of the four story stem factors grouped by four-way
classification of mother's adult attachment interview
Table 1:
Factor
Means (SID)
Open/positive
df
f-ratio
dismissing
(n=20)
autonomous
(n=50)
preoccupied
(n= 12)
unresolved
(n=7)
1.7503 (. 20)
1.8000 (. 25)
1.6856 (. 35)
1.8333 (. 11)
3,85
84
.
47
.
1402(. 07)
1847(. 10)
1295(. 06)
1580(. 10)
3,82
1.64
18
response
Discipl ine/punishment
Positive maternal
representation
Note:
f-prob
5650(. 24)
.
5118(. 17)
.
5890(. 15)
.
6133 (. 21)
.
3,8 5
1.03
1295(. 11)
1203(. 12)
1364(. 11)
1169(. 11)
3,82
07
.
-38
97
.
based
p values are
on two-tailed levels of significance
Means and SDS of the four story stem factors grouped by four-way
fnther'sndult
attachment interview.
(+q,,-,1fir,qtic)n nf
Table 2:
Factor
df
Means (SD)
f-ratio
dismissing
(n= 18)
autonomous
(n=57)
preoccupied
(n= 10)
unresolved
(n=3)
1.6594 (. 35)
1.7999 (. 22)
1.8253 (. 15)
1.8771 (. 13)
33,8
4
1.76
1828(. 10)
1553)(. 08)
2030(. 13)
.
1515 (. 12)
81
-3),
89
.
44
.
86
.
Disc ipl ine/punishment
Control I Ing/negative
zn
5703 (. 2 1)
.
5416(. 19)
.
5121 (. 19)
.
4970(. 21)
.
33,8
4
24
.
1591 (. 13)
1274(. 11)
0864(. 08)
.
0303(. 02)
.
33,8
1
1.14
Positive maternal
representation
f-prob
16
354
APPENDI
Table 1: Means and SDS of two-way interaction of gender and insecure vs

secure
attachment classification with mother at 12 months for the r)erformance vari2h]f-.,,
Variables
Means (SDS)
insecure
F-value
(df)
secure
girls
(n= 17)
boys
(n= 2 1)
girls
(n= 23)
boys
(n= 27)
Role of parent
1.7861 (. 39)
1.6147 (. 52)
1.8498 (. 38)
1.8855 (. 45)
1.15(1,87)
28
.
Child's understanding
of conflict
1.0428 (. 14)
1.0260 (. 16)
1.0830 (. 09)
1.0404 (. 19)
14(1,87)
70
.
Directness of
performance style
1.8610 (. 26)
1.7619 (. 40)
1.9763 (. 05)
1.9428 (. 20)
34 (1,87)
55
.
Responsivity to
examiner
2.5615
2.6320
2.8221
(. 29)
2.7643 (. 44)
4-3)(1,87)
.
.51
Involvement of
examiner
1.4652 (. 33)
1.4675 (. 29)
1.4032 (. 23)
1.4444 (. 29)
09(1,87)
.
75
.
Investment in
performance
2.1872
(. 80)
2.1558 (. 67)
2.2055 (. 64)
2.2189 (. 68)
02(1,87)
.
88
.
2620 (. 18)
.
3463 (. 20)
.
2490 (. 15)
.
2694 (. 13)
.
76(1,87)
.
38
.
Adaptiveness of
response
1.7540 (. 43)
1.4935 (. 27)
1.9012 (. 42)
1.6768 (. 33)
05(1,87)
.
82
.
Narrative coherence
2.6203
(. 46)
2.6902 (. 55)
26(1,87)
.
60
.
1833 (. 30)
0540 (. 13)
.
0853 (. 13)
.
01 (1,87)
.
89
.
Joy
4403
.
(. 26)
4906 (. 22)
.
5468 (. 25)
.
5230 (. 25)
.
46(1,87)
.
49
.
Anger
0606 (. 08)
.
0476 (. 05)
.
0817 (. 05)
.
0404 (. 04)
.
1.29(1,87)
25
.
Distress
0196
.
(. 01)
0202 (. 03)
.
0316 (. 04)
.
0258 (. 03)
.
12(1,87)
72
.
Concern
0553 (. 10)
.
0332
.
0382 (. 04)
.
0494 (. 06)
.
63328 (. 2 7)
.
5859 (. 32)
.
6219 (. 26)
.
5556 (. 26)
Denial
Control
I Anxiety
(. 55)
(. 50)
1408 (. 20)
2.5931
.
(. 50)
2.8340
(. 55)
(W) I
1
1.33)(1,87)
1
02(1,87)
.
,I
-87
.
55
APPENDIX F. 1
Coding sheetfor predicting which attachmentclassification will score high vs low (present
or absent)for
each variable
Content Variables
Si
VARIABLE
aff
comply
dishon
h
I
guilt/repar
icr
inj
ncomp
h
I
negatyp
oexcl
phy agg
p empathy
h
I
posatyp
h
I
phy pun
p repair
self exclude
h
I
sharneother
h
I
v pun
h
I
MH
3C
BH
LK
SS
SN
RN
BS
FX
BT
356
APPENDIX G. 1 (comt)
Parental Representation variables
SJ
VARIABLE
disc f
MH
3C
BH
LK
SS
SN
RN
BS
EX
BT
h
I
-I
disc m
h
I
f
neg
h
I
neg m
h
I
f
pos
pos m
Performance code variables

si
VARIABLE
adr
MH
h
I
-
cuc
h
I
di
h
I
dni
le
ip
nc
res
rp
3C
BH
LK
SS
SN
RN
BS
EX
BT
57
APPENDIX (j. 1_Lcont)

Affect variables
VARJABLE
si
pang
nang
tang
panx
nanx
tanx
pcm
ncrn
tcrn
pctr
nctr
tctr
pdis
h
I
ndis
h
I
tdis
PiOY
nj oy
11 10\
j ,-I
Secure
-B
--Resistant
-C
MH
3C
BH
LK
SS
SN
RN
BS
Avoidant -A
Disorganised -D
EX
BT
358
APPENDIX
Discriminant
function analysis
Predicting groups by infant security with mother, using the four factor
scalesand the
discriminating variable of father's social class.
Canonical
discriminant
Group
Func
1
2
functions
evaluated
at
group
means
centroids)
47761
-.
34826
.
Classification
results
Membership
12
--------
--------
------
--------------------
Group
Predicted
No. of
Cases
Group
Actual
Group
36
15
41.7%
21
58.3%
Group
49
15
30.6%
34
69.4%
Ungrouped
10
cases
.
Percent
(group
of
Classification
"grouped"
cases
processing
correctly
0%
cla ssified:
1
100.0%
57-65%
summary
106 (Unweighted)
cases
0 cases
were excluded
had at least
20 cases
86 (Unweighted)
cases
were processed.
group
for missing
or out-of-range
discriminating
variable.
missing
one
for
output.
printed
were used
codes.
359
APPF. NIIIY
r- i
Predictinggroupsby infant security

with mother, using the four factor scales
calculated
per story and the discriminating variable of father's
social class.
discriminant
Canonical
functions
Func
Group
1
2
at
group
results
No. of
Cases
------
Predicted
12
--------
Group
Membership
--------
Group
38
17
44.7%
21
55.3%
Group
50
10
20.0%
40
80.0%
10
cases
.
of
Classification
centroids)
--------------------
Percent
(group
Group
Ungrouped
means
44080
-.
32142
.
Classification
Actual
evaluated
"grouped"
cases
processing
correctly
0%
cla ssified:
1
100.0%
64.77%
summary
106 (Unweighted)
cases
0 cases
were excluded
17 cases
had at least
89 (Unweighted)
cases
were processed.
for missing
group
or out-of-range
discriminating
variable.
one missing
for
output.
printed
used
were
codes.
360
APPENDIX
Predicting groups by infant security with father,
using the four factor scales calculated
per story and the discriminating variable of father's social class.
discriminant
Canonical
functions
Func
Group
1
2
results
No. of
Cases
0%
cases
Classification
0%
63
.
of
centroids)
"grouped"
cases
processing
Membership
--------
Percent
(group
12
25
Group
Predicted
--------
------
Ungrouped
means
--------------------
Group
group
Group
Group
at
37419
.
14700
-.
Classification
Actual
evaluated
correctly
0%
classified:
25
100.0%
63
100.0%
11
100.0%
71.59%
summary
106 (Unweighted)
cases
0 cases
were excluded
had at least
7 cases
99 (Unweighted)
cases
were processed.
for missing
group
or out-of-range
discriminating
variable.
missing
one
for
output.
printed
were used
codes.
361
APPENDIX
G. 1
Predictinggroupsby maternalsecurityof attachment,usingthe four factor scales the

and
discriminating variable of father's social class.
discriminant
Canonical
Group
functions
group
results
No. of
Cases
Group
------
Predicted
Group
12
--------
--------
35
9
25.7%
26
74.3%
Group
51
9
17.6%
42
82.4%
Classification
106
0
20
86
centroids)
Membership
Group
of
(group
--------------------
Percent
means
30342
-.
20633
.
Classification
Actual
at
Func
1
2
evaluated
"grouped"
cases
processing
correctly
classified:
59.30%
summary
(Unweighted)
cases
cases
were excluded
had at least
cases
(Unweighted)
cases
were processed.
for missing
group
or out-of-range
discriminating
variable.
one missing
output.
were used for printed
codes.
362
APPFNDIV
(;. I
Predicting groups by
maternal security of attachment,
four factor
the
using
scales
calculated per story and the discriminating
variable of father's social class.
discriminant
Canonical
evaluated
at
group
means
centroids)
-. 55659
37848
.
1
2
Classification
results
No.
of
Cases
Group
Predicted
12
Group
Membership
--------
-------Group
35
15
42.9%
20
57.1%
Group
51
6
11.8%
45
88.2%
Percent of
Classification
106
0
20
86
(group
Func
Group
Actual
functions
"grouped"
cases
Processing
correctly
classified:
69.77%
summary
(Unweighted)
cases
cases were excluded
had at least
cases
(Unweighted)
cases
were processed.
group
for missing
or out-of-range
discriminating
variable.
one missing
output.
codes.
363
APPENDIX G. 1
Predicting groups by paternal security of attachment,
using the four factor scales
calculated per story and the discriminating variable of father's social class.
Canonical
discriminant
Func
Group
evaluated
at
group
Classification
results
centroids)
No. of
Cases
Group
Predicted
12
Group
Membership
--------
Group
27
2
7.4%
25
92.6%
Group
58
2
3.4%
56
96.6%
cases
Percent of
Classification
106
0
20
86
(group
--------
Ungrouped
means
32077
.
15466
-.
1
2
Actual
functions
"grouped"
cases
processing
correctly
01
0%
classified:
100.0%
68.24%
summary
(Unweighted)
cases
cases were excluded
cases had at least
(Unweighted)
cases
were processed.
for missing
group
or out-of-range
discriminating
variable.
missing
one
output.
codes.
(LOINDON)
\! M. /
364
APPENDIX
G. 2
Cluster analysis using the four factors (Quality/Open Response,Discipline/Punishment,

Controlling/Negative and Positive Maternal Representation) as
selectedcharacteristics.
Rescaled
CASE0
Label
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
45
79
62
26
42
7
9
100
52
19
54
69
86
50
99
18
70
81
23
68
61
65
72
71
24
101
6
53
77
80
15
Case 105
Case
Case
Case
Case
Case
Case
case
case
case
Case
Case
Case
Case
Case
Case
Case
case
Case
Case
Case
case
Case
Case
Case
Case
Case
Case
case
Case
case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
Case
63
14
27
92
10
56
13
29
41
47
Num
45
79
62
26
42
7
9
100
52
19
54
69
86
50
99
18
70
81
23
68
61
65
72
71
24
101
6
53
77
80
15
105
63
14
27
92
10
56
13
29
41
47
89
89
97
97
2
2
84
84
36
36
64
64
57
57
35
35
51
51
37
37
21
21
67
67
104 104
67
87
93
93
40
40
66
66
33
33
39
39
102
102
95
95
5
5
34
34
78
78
16
16
73
73
22
22
17
17
55
55
74
74
6
8
59
59
60
60
49
49
75
75
31
31
32
32
48
48
25
25
96
96
76
76
11
11
43
43
46
46
+-
Distance
5
10
------------------------------------
Cluster
Combine
15
20
25

Mac Arthur Story Stem Battery

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Mac Arthur Story Stem Battery

Hochgeladen von

Copyright:

Verfügbare Formate

THE REFLECTION

Susan Anne Yabsley

Submitted for the degree of Doctor of Philosophy

This series of studies reports on the prospective and concurrent relation of

Assessing the Internal World of the Young Child

Analysis of the MacArthur Story Stem Battery Narratives

The Effects of Demographic Factors on the Children's Responses to

Associations between the Child Behaviour Checklist and the

Associations Between the MacArthur Story Stem Battery and

9.3.1 Gender and infant-mother attachment

Mill Hill Vocabulary Scale

Child Behaviour Checklist

Interaction of gender and attachment

Coding Sheet for Profile

Median Cohen's Kappa scores for the inter-rater reliability of the 24

Median Kappa scores for inter-rater reliability of the complex content

Median Kappa scoresfor inter-rater reliability of the new variables derived

Median inter-rater reliability for the parental representations manual

Median inter-rater reliability for mother and father representations 75

Median inter-rater agreement for the performance codes as presented

Median scores for inter-rater reliability of the complex performance

Median inter-rater reliability for narrative coherence

Median inter-rater reliability for the affect themes as presented in the

Median inter-rater reliability of the complex affect themes collapsed to

Median inter-rater reliability of affect themes when three phases of the

Reliable content codes and aggregate alphas across the II stories

Parental representations and their aggregate alphas across storles

Sorted factor loadings for the unrotated solution

Sorted factor loadings for the varimax rotation

Means, standard deviation, range and Cronbach's alpha per story by

Means,, standard deviation, range and Cronbach's alpha per story by

Intercorrelations between the eleven story stems for factor 1, quality/open

Presentsthe Intercorrelations between the eleven story stems for factor 2.

Bivariate correlations and levels of significance of mother's, father's and

Correlations and levels of significance for parents' expressive language

Correlations and levels of significance of infant temperament and the story

Means and SDS of factor scores grouped by three classifications of

Means and SDSs between factor scores grouped by three classifications

Means, SDS and levels of significance for the performance variables by

Correlations of mother's and father's rating on CBCL scales

Correlations of mother's rating of the child on the intemalising scale with

Correlations of mother's ratings of the child on the intemalising scale with

Correlations of mother's rating of the child on the externalising scale with

Correlations of physical aggressionand the extemalising scale as rated by

Correlations and levels of significance of father's rating of the child on the

Correlations of father's rating of the child on the extemalising scale with

Correlations of physical aggressionand the externallsing scale as rated by

Infant-parent security grouped by instrument. forced classlficatlon in

Means and SD of father's social class and security of attachment to mother

Means and SIDS of story stem factors grouped by original three-way

Mean rank of the Discipline/Punishment factor grouped by three-way

Means, SDS of story stem factors by three-way classification of

Means and SDS of insecure vs secure attachment with mother at 12

Means and SDS of secure vs insecure attachment with mother at 12

Means and SDS of performance variables grouped by insecure vs secure

Means and SDS of securevs insecure attachment with father at 18 months

Means and level of significance for the performance variables by insecure

Means and SD of father's social class and mother's AAI attachment

Means and SDS of story stem factors grouped by three-way classifications

Means and SD of father's social class and father's AAI attachment

Means and SDS of story stem factors grouped by three-way classification

Means and SDS of the two-way interaction of gender and insecure vs

Means and SDS of two-way interaction of gender and insecure vs secure

Means and SDS of the two-way interaction of gender and insecure vs