MICHIGAN STAT

W Ill/xfl/il/“ii/li/‘ii/ll/l‘llli/i

3 ‘2 3 1050 3286

Hm;

This is to certify that the

dissertation entitled

Early Childhood Trauma and Maternal Childrearing
Attitudes: Their Contributions to Object Representations
in Delinquent and Nondelinquent Adolescent Boys

presented by

Sandi Lyn Isaacson

has been accepted towards fulfillment
of the requirements for

 

 

 

 

Ph.D. degree in Counseling Psychology
Major professor I /
Date
MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or betore date due.

DATE DUE DATE DUE DATE DUE

     

  
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Affirmative Action/Equal Opportunity Inetitution
Wane-m

l_ f;...l_._.#____,_—.——_.

 

 

EARLY CHILDHOOD TRAUMA AND MATERNAL CHILDREARING ATTITUDES:
THEIR CONTRIBUTIONS TO OBJECT REPRESENTATIONS IN DELINQUENT
AND NONDELINQUENT ADOLESCENT BOYS
By

Sandi Lyn Isaacson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Counseling, Educational Psychology
and Special Education

1996

ABSTRACT
EARLY CHILDHOOD TRAUMA AND MATERNAL CHILDREARING ATTITUDES:
THEIR CONTRIBUTIONS TO OBJECT REPRESENTATIONS IN DELINQUENT
AND NONDELINQUENT ADOLESCENT BOYS
By

Sandi Lyn Isaacson

Juvenile delinquency has become a major economic and sociological problem.
Delinquent youth are a perplexing, treatment-resistant population who present mental
health professionals with increasing demands for services. Prevailing theories have been
unable to explain the interrelationships among the numerous etiological factors found to
contribute to delinquency. Attachment theory postulates that based upon experiences in
early childhood, particularly relationships with early caregivers, individuals develop
internal representations of themselves and others which organize subsequent interpersonal
behavior. Those individuals who experience parenting which does not adequately meet
their emotional or physical needs, or who experience early childhood trauma develop
internal representations of others as hurtful and unavailable and are likely to react with
hostility and aggression toward others.

This study utilizes attachment theory to understand previous findings and
investigates the relative contributions of early childhood trauma and maternal childrean'ng
attitudes to the object representations of delinquent and nondelinquent adolescent boys as
measured by the Rorschach Inkblot Test. Delinquent boys experienced more frequent and
more severe early childhood trauma than nondelinquent boys. Within the subgoup of

majority boys, delinquent boys’ mothers endorsed more harsh, abusive attitudes towards

childrearing; for the subgroup of minority boys, nondelinquent boys’ mothers endorsed the
harsher childrearing attitudes. No significant differences were found between delinquent
and nondelinquent boys on Rorschach variables which reflect object representations or on
a measure of boys’ perceptions of their attachment relationships with their mothers.
Observed results from the Rorschach Inkblot Test suggest that urban adolescents in
general are more alienated and detached than previously thought. Nondelinquent
adolescents were found to have better verbal skills, more responsive mothers, and to have

experienced less early trauma, factors which may prevent engagement in delinquent

behavior.

Copyrighted by:
SANDI LYN ISAACSON
1996

To those who shared the dream.

ACKNOWLEDGMENTS

Many have commented that completing a dissertation is a lonely process. This has not
been true for me. I have had the good fortune to have many fiiends, colleagues and mentors
support and guide me throughout this process. While it would be impossible to thank each of
them individually, I would like to recognize those whose help has been invaluable.

I deeply appreciate the support and guidance of my chairman, Dr. Frederick G. Lopez,
and committee members, Dr. W. Patrick Dickson, Dr. Elizabeth A Seagull, and Dr. Robbie J.
Steward. Each basicontributed in a unique and profound way to my personal and professional
growth.

The assistance and cooperation of the Ingham County Probate Court and The
Honorable Donald S. Owens, Chief Judge as well as the Ingham County Youth Center, its
superintendent, Mr. Nervy C. Oliver, and psychologist, Dr. Garrett L. Turke, inspired this
study and made it possible. It is through Dr. Turke’s mentorship and compassionate
supervision that I first became interested in the area of delinquency and learned how to “hang in
there” with myself and others. I will be forever indebted to Dr. Turke and Mr. Oliver for their
dedication, support, friendship, and unwavering confidence in me.

I would like to thank the LanSing School District, Dr. Richard Halik superintendent,

the Ofiice of Student Support Services and the Ofice of Research and Evaluation Services and

the Boys and Girls Club of Greater Lansing for assisting with data collection and for allowing
me access to their facilities.

A special thank you to Grace Gibson for her ceaseless dedication and countless hours
of work that went well above and beyond anything I had hoped for. Grace began this project
as my research assistant and concluded it as my friend, the most important outcome measure of
all.

I would also like to extend a heartfelt thank you to all of my friends who helped to
keep me “ uman” during the toughest of times. To Stuart, Richard, Diane, Deb, Jim, Linda,
Steve, Jerry, Tom, Frank, Terry, and the rest of the gang, thanks for always believing in me and
never giving up.

Finally, I would like to thank the boys and their mothers who participated in this study.
I recognize that it was not always easy for you to share so honestly with a stranger. Many of
you expressed the hope that this work would somehow help another youngster choose a better

path in life. I share this hope with you.

vii

TABLE OF CONTENTS

List of Tables .

INTRODUCTION

Introduction
Purpose
Theoretical Framework
Attachment Theory
Object Representations
Early Life Events
Parental Attitudes Toward Childrearing
Life Events and Parental Childrearing Attitudes
Perceptions of Parents as Object Representations
Hypotheses

LITERATURE REVIEW . .
An Attachment Theory Perspective on Juvenile Delinquency.

From Working Models to Object Representations
Delinquency and Traumatic Life Events
Parenting Attitudes and Delinquency.

Object Representations of Delinquents

Delinquency and Perceptions of Parent-Child Relationships .

Summary

METHODOLOGY

Participants
Procedure
Measures . .
Children’s Life Events Scale .
Adult-Adolescent Parenting Inventory
Rorschach Inkblot Test
Human Content Responses (H)
Human Movement Responses (M)
Texture Responses (T, TF, FT)
Object Representations and the Rorschach
Delinquents and the Rorschach

viii

y—ey—e
t-‘o

X

WQQMWWWI—hu—I

.12
.16
.19
.20
.25
.26
.28

.30
.30
.32
.35
.35
.37
.38
.39
.40
.41
.42
.44

Inventory of Parent-Peer Attachment .

Peabody Picture Vocabulary Test - Revised .

Research Hypotheses .
Analysis
4. RESULTS . .
Demographic Variables
Results of Hypothesis Testing.
Traumatic Life Events.
Maternal Childrearing Attitudes
Object Representations
5. DISCUSSION .
Summary . . . .
Implications for Theory and Practice .
Limitations of the Present Study
Recommendations for Future Research
APPENDICES
A. Consent Letters
B. General Information Questionnaire
C. Children’s Life Events Scale - Preschool Version
D. Inventory of Parent—Peer Attachment .
REFERENCES

ix

.45
.47
.49
.51

.53
.53
.59
.59
.61
.63

.68
.68
.76
.78
.83
.86
.86
.88
.89
.91

.94

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

Table 8

LIST OF TABLES

Distribution of Demographic Variables

Descriptive Statistics of Dependent Variables

Correlation Matrix for Dependent and Independent Variables
Analyses of Variance for Delinquents and Nondelinquents

Descriptive Statistics and Intercorrelations of Maternal
Childrearing Scales

Summary of Hierarchical Regression Analysis for Variables
Predicting Object Representations . .

Summary of Hierarchical Regression Analysis for Variables
Predicting IPPA scores . . .

Comparison of Means and Frequencies Among Groups of
Adolescents on Select Rorschach Variables

.54

.56

.58

.60

.62

.66

.74

CHAPTER]

Introduction

Juvenile delinquency has become a major economic and social problem. In 1989,
youth between ages 13 and 18 years old accounted for 18.7% of arrests while comprising only
8.3% of the population (Maguire & Flanagan, 1991). In 1994, they accounted for 21.2% of all
arrests while still comprising only 8.3% of the population (Federal Bureau of Investigation,
1995). Between 1989 and 1994 the arrest rate of juveniles under age 18 increased 172% for
murder and non-negligent manslaughter, 126% for forcible rape, and 180% for aggravated
assault while the rates for adults remained virtually unchanged (Federal Bureau of
Investigation, 1995).

Juvenile delinquents are a perplexing, treatment-resistant population who often appear
hostile, angry, impulsive, rnistrustfiIl, and unable to form useful therapeutic alliances (Willock,
1986, 1987). Biological, sociological and psychological theories have been proposed to
explain the nature and etiology of delinquent behavior. The prevailing theories, however, have
not addressed the possible interrelationships among early childhood experiences, parental
childrearing attitudes, subsequent object representations and their conjoint contributions to

delinquency.

Attachment theory has recently been suggested as a fi'amework for linking early
childhood experiences to subsequent delinquent behavior in adolescents (Allen, Aber, &
Leadbeater, 1990; Deutsch & Erickson, 1989; Weber, 1990). Attachment theory, which
proposes that humans have an innate biological need to form meaningful interpersonal
relationships (Bowlby, 1958, 1977a, 1977b), ofi‘ers a viable explanation for the development,
continuation and cessation of delinquent behavior.

Juvenile delinquency has been related to a history of frequent, traumatic life events
prior to four years of age (Deutsch & Erickson, 1989). It has also been related to poor family
functioning (Loeber & Dishion, 1983; Loeber & Stouthamer—Loeber, 1986) and various
inadequacies in parenting such as parental inconsistency and emotional unavailability (Brand,
Crous, & Hanekom, 1990; Davidow & Bruhn, 1990; F eehan, McGee, Stanton, & Silva, 1991;
Steinberg, 1987; Wells & Rankin, 1988), lack of nurturing and protection (Davidow & Bruhn,
1990) and familial violence and abuse (Lewis, Shanoclg Pincus, & Glaser, 1980; Steinberg,
1987).

Research has not yet integrated these distinct findings into a more comprehensive
model of delinquent behavior. Attachment theory provides a framework for hypothesizing that
parental attitudes moderate the effects of stressful events occurring in early childhood to the
development of object representations. Object representations, which are internal models of
others and influence interpersonal behavior, are poorly developed in individuals with conduct
disorders and antisocial personality disorders. These individuals tend to be detached fiom
others and avoid engaging in close interpersonal relationships (Gacono & Meloy, 1991; 1994;

Meloy, 1998; Weber, 1990; Weber, Meloy, & Gacono, 1992). Distortions in the perception of

self and others have been found in youngsters who commit a wide range of criminal acts
(Exner, 1986; Greco & Cornell, 1992; Ray, 1963), suggesting that delinquency may also be
related to poorly developed object representations. Further research is needed to clarify the
contributions of life events and parental childrearing attitudes on object representations for

delinquent and nondelinquent adolescents.

Luring
The purpose of this study was to examine differences between delinquent and
nondelinquent boys in their experiences of early childhood life events, in their mothers’
attitudes toward childrearing, and in their current object representations. The study fiIrther
investigated whether maternal attitudes toward childrearing moderate the efl‘ect of eariy
childhood life events upon current object representations. Before identifying the specific 7‘
hypotheses under investigation, the remaining sections of this chapter will briefly overview the

theoretical fiamework guiding this investigation and highlight some key findings fiom the

Q
existent literature. A more detailed discussion of the literature will be found in Chapter 2.

Theoretical Framework

 

Attachment Theory

Attachment theory posits that parent-child interactions in infancy and early childhood
(a) form the foundation for the way in which individuals perceive themselves in relationships
with others and (b) organize how individuals behave in interpersonal situations. Children who

have formed “secure” attachments are more likely to view the world as a nurturing,

trustworthy place and to enjoy engaging in emotionally close, reciprocal interpersonal
relationships. Children who do not form secure attachment relationships with their caregivers,
on the other hand, are likely to experience significant dificulty with interpersonal relationships
throughout their lifespan (Ainsworth, 1989; Bowlby, 1977a; 1977b; 1988; Goldberg, 1991;
Paterson & Moran, 1988).

Those persons whose parents are not adequately responsive in infancy are more likely
to develop insecure attachments (Bowlby, 1988; Cicchetti & Barnett, 1992; Main & Stadtrnan,
1981; Spieker, 1986). They are more likely to experience behavioral difliculties in childhood
and evidence various forms of psychopathology in adulthood (Bowlby, 1977a; Cicchetti &
Barnett, 1992; Lyons-Ruth, Alpem, & Repacholi, 1993; Parkes, 1982; Paterson & Moran,
1988). These insecurely attached individuals generally perceive other people as hostile,
unavailable, and unable to meet their needs. They tend to be more aggressive, oppositional,
and violent in childhood and adolescence (Cohn, 1990; Lyons-Ruth et al., 1993; Renken,
Egeland, Marvinney, Mangelsdorfi & Sroufe, 1989; Speltz, Greenberg, & Deklyen, 1990;
Sroufe, 1982).

Adolescents with insecure attachments to parents are at an increased risk of engaging
in delinquent and other maladaptive behavior (Allen et al., 1990). Members of this group are
more likely to evidence conduct problems, inattentiveness and depression (Nada Raja, McGee,
& Stanton, 1992; Burbach, Kashani, & Rosenberg, 1989). Adolescent aggression can be
predicted fiom a prior history of impaired parent-child relations (Haapasalo & Tremblay, 1994;

Sirnons, Robertson, & Downs, 1989).

Object Represent_a_tro_n_s

A child's perception of human relationships is influenced by the interaction between
parent and child throughout the child's early years. This perception ultimately becomes a stable
model for representing all interpersonal relationships (Homer, 1984; Mayrnan, 1967).
According to Bowlby, the originator of attachment theory, this representation becomes
internalized by approximately five years of age (Bowlby, 1977a). Based upon experiences of
eliciting or failing to elicit caring and protection fiom a primary caregiver during these early
years, individuals develop “internalized working models” of themselves and others (Bowlby,
1958; Crittenden, 1990; Hopkins, 1990).

Bowlby’s “internal working models” are similar in structure and firnction to early
“object representations” of object relations theory (Homer, 1984,1991; Levine & Tuber,
1993). Both reflect the relationship with a primary (usually maternal) caregiver and, once
formed, exert a persistent, unconscious influence upon cognition, affect and behavior (Blatt &
Lerner, 1983; Bowlby, 1977a, 1988; Homer, 1984; Maymarr, 1967; Stricker & Healey, 1990).

Conduct-disordered youngsters and psychopathic adults have been noted to have
poorly developed object representations (Exner, 1986; Gacono, 1988; 1990; Gacono & Meloy,
1991; Meloy, 1988; Weber, 1990; Weber, Meloy, & Gacono, 1992). Delinquent youth have
also been found to have impaired object representations (Exner, 1986; Exner, Bryant, & Miller,
1975; Greco & Cornell, 1992; Ray, 1963). This study proposed that serious delinquents, those
who have committed acts that would be felonies if committed by an adult, would differ
significantly from nondelinquents in the quality of their object representations. It firrther

proposed that these differences were related to measurable diflerences in their history of

stressful life events and differences in parental childrearing attitudes between mothers of

delinquent and nondelinquent boys.

Early Life Events

Research has related the severity of delinquent behavior to the amount of trauma
experienced in early childhood (Deutsch & Erickson, 1989). An early study by Bowlby (1947)
related juvenile theft to early maternal loss and fiequent changes in caregivers. More recently,
studies have shown that psychological “loss” of the parent due to the parent’s emotional
unavailability may be just as damaging as actual parental loss (Bowlby, 1984).

Children’s physical, emotional and behavioral adjustment problems have been
associated with environmental and familial stressors including parents’ physical health and
hospitalimtion, financial deprivation, and parental divorce (Longfellow & Belle, 1984; Loss,
Beck, & Wallace, 1995). Children and adolescents hospitalized for physical ailments were
found to experience a greater number of stressfirl life events in the year prior to hospitalization
than their non-hospitalized peers (Coddington, 1972a, 1972b, 1984). This suggests that
stressful life events contribute to physical illness.

Stressful events experienced in early childhood can also affect emotional and behavioral
development. Young children identified by their teachers as having adjustment problems
experienced Significantly more stressful life events in the preceding year than children who did
not display such problems (Sandler & Block, 1979). Undersocialized delinquent adolescents
experienced a greater number of environmental stressors, especially between the ages of two

and four years old, than their socialized delinquent peers (Deutsch & Erickson, 1989).

While the impact of certain life stressors such as parental criminality or broken homes
upon delinquent behavior has received a significant amount of attention, there is a need to
examine the impact of a wide range of stressful life events upon delinquency, especially events
occurring in early childhood (Deutsch & Erickson, 1989; Kolvin, Miller, Fleeting, & Kolvin,
1988). This study examined differences in the occurrence of a wide range of stressfirl events
using a comparison group of nondelinquent youth fi'om the same geographic and

socioeconomic background and focusing on life events which occurred prior to four years of

age.

Mal Attitudes Toward Chfimaflg

Maltreated children are more likely to engage in delinquent and criminal behavior as
adolescents and adults (Luntz & VVIdom, 1994; Van Voorhis, Cullen, Mathers, & Gamer,
1988; VVIdom, 1989a). Maltreatment has been strongly associated with specific parental
attitudes towards child rearing which include lack of empathy, inappropriate developmental
expectations, role reversal, and endorsement of physical punishment (Call, 1984).
Investigations of child rearing attitudes of parents of delinquent youth have reported similar
findings.

The parents’ interest in the child and the management strategies they use to influence
the child’s behavior are among the most powerful predictors of delinquency (Loeber &
Dishion, 1983; Loeber & Stouthamer-Loeber, 1986; Snyder & Patterson, 1987). Parental
rejection, indifference and belief in harsh physical punishment have all been associated with the

development of delinquent behavior (F anington, 1978; Gove & Cnrtchfield, 1982; Haapasalo

& Tremblay, 1994; Sirnons, Robertson, & Downs, 1989). Children whose parents are unable
to respond empathically are especially prone to violent, aggressive behavior (F unnan &
Funnan, 1984; Sroufe, 1982).

Studies that directly observed parental dynamics indicated that maternal hostility,
inconsistent discipline, and rejection are all strongly related to adolescent delinquency (DiLalla,
Mitchell, Arthur, & Pagliocca, 1988; Farrington, 1978; Feehan et al., 1991; Renken et al.,
1980; Snyder & Patterson, 1987). Family dynamics such as expressions of caring, trust,
appropriate supervision and communication are stronger predictors of delinquency than the
effects of family structure (Cemkovich & Giordano, 1987).

It often remains unclear whether the aberrant childrearing practices associated with
parents of delinquent adolescents developed in response to raising a troubled child or are a
reflection of the parent's fundamental attitudes toward the task of childrearing (Liska & Reed,
1985). This study examined mothers’ overarching attitudes toward childrearing by
administering a well established measure of parenting attitudes and beliefs to the mothers of

delinquent and nondelinquent youth.

Life Eventsi and Parental Child Rearing Attitudes

According to attachment theory, “the roles of the caregiver are first to be available and
responsive as and when wanted and, secondly, to intervene judiciously should the child or older
person who is being cared for be headed for trouble (Bowlby, 1977, p. 204).” A child may
become insecurely attached if he or she experiences traumatic events to which the parents’

response is inadequate. Parents who are less empathic toward their children or have unrealistic

expectations such as expecting that children will satisfy parental emotional needs are less likely
to be appropriately responsive in the face of stressful events.

In an efl‘ort to summarize the multiple contributing factors to delinquent behavior,
Lewis (1992) states:

...whatever increases irnpulsivity and irritability, engenders hypervigilance and

paranoia, diminishes judgment and verbal competence, and curtails the ability

to recognize one's own pain and the pain of others, also enhances the tendency

toward violence. (p. 388)

It is no surprise that risk of delinquency increases in the presence of multiple family risk
factors. Studies which examine life stressors as well as family dynamics find that both
contribute to delinquency (Haapasalo & Tremblay, 1994; Kolvin et al., 1988; Renken et al.,
1989). These studies do not comment, however, on how stressors and family dynamics may
interact to promote delinquency.

Deutsch and Erickson (1989) suggested that parental responsiveness moderates the
effect of traumatic life events on child adjustment. Maternal well-being and childrearing
practices have been found to buffer the impact of economic and situational stressors on
children's adjustment problems (Dodge, Pettit, & Bates, 1994; Longfellow & Bell, 1984).
Parenting process variables have also been found to mediate the efl‘ect of economic poverty on
delinquency (Sampson & Lamb, 1994).

This study proposed that matemal childrearing attitudes moderate the efl’ect of life

events on delinquency primarily by way of their influence on the development of object

10

representations. This investigation compared the relative contributions of stressful life events

and maternal childrearing attitudes toward delinquency.

Perceptions of Pgentsfias Object RernesenMs

Conduct disordered and delinquent youth in treatmmt settings perceive their parents to
be more controlling, rejecting, and inconsistent, and less caring and afi’ectionate than their peers
(Brand et al., 1990; Burbach et al., 1989; Cemkovich & Giordano, 1987). Delinquents report
that their earliest memories of their parents involve instances of parental unavailability, inflicting
physical injury, and responding with a lack of nurturing and protection (Davidow & Bmhn,
1990). Delinquent adolescents report a history of inconsistent parenting with parental figures
unpredictably alternating between harsh, punitive physical punishment and lax enforcement of
rules (Brand et al., 1990; Haapasalo & Tremblay, 1994; Van Voorhis et al., 1988). The
perception of parental rejection is one of the strongest predictors of firture delinquency in self-
report studies (Sirnons et al., 1989). These responses may be an artifact of the security present
in treatment settings where such afi'ectively honest responses can be tolerated.

In untreated delinquent and conduct disordered youth, the continued need to defend
against feelings toward the attachment object may lead to incongruent, idealized attitudes
toward the attachment figure on measures which tap conscious attitudes. It is unclear whether
the responses of delinquent youth incarcerated in a detention facility will reflect defensive,
idealized attitudes toward attachment figures or the more afl’ectively candid responses of youth

in treatment facilities that have been the subject of prior investigations.

11

Hypotheses

Delinquent males have experienced a higher frequency of stressfiIl life events in early

childhood than nondelinquent males.

Delinquent males have experienced stressful life events in early childhood that are more

traumatic than the stressfirl life events experienced by nondelinquent males.

Mothers of delinquent males have less empathic and more controlling attitudes toward

childrearing than do mothers of nondelinquent males.

Compared to their nondelinquent peers, delinquent males have impaired object

representations.

Maternal attitudes toward child rearing moderate the efl’ect of early childhood life

events upon internal object representations.

CHAPTER 2

Literature Review

This chapter presents an overview of attachment theory which provides the underlying
framework for this study. The literature regarding the impact of stressfirl early childhood
events and parental child rearing attitudes on the development of attachment relationships and
the salience of attachment relationships to the formation of internal object representations is
reviewed. The contribution of these factors to understanding the nature of delinquency is also

reviewed.

An Attachment Theory Perspective on Juvegle Delinquency

Attachment theory emphasizes the salience of close relational bonds or “attachments”
in the development of animals and humans. The primary firnction of attachments is to protect
the young from danger by maintaining close proximity to a larger, stronger attachment figure
who can fend off attacks frOm predators (Bowlby, 1958). For humans, the function of
attachment relationships is to provide a perceived sense of security which includes an internal
feeling of safety as well as an actual protection from physical harm (Sroufe & Waters, 1977).
This sense of security is achieved by establishing and maintaining a relationship with a primary

attachment figure, generally a mother.

12

13

Classical psychoanalytic theory proposes that the infant’s attachment to his or her
mother results fi'om the experience of being fed by her (Freud, [1905] 1953). Attachment
theory differs by postulating that the formation of attachments is a biologically determined,
primary drive that develops independently of oral drives (Ainsworth, 1969). The infant is born
with a variety of instinctual behaviors which facilitate the development of an attachment
relationship between infant and mother. Conversely, the mother is also biologically primed to
respond to these behaviors in ways that enhance the relationship. The attachment systems of
mother and infant are essential for normal emotional development (Kestenbaum, 1984).

The primary attachment relationship forms the template for fiIture interpersonal
relationships. This early mother-child relationship organizes the child’s firture behavior
towards others (Ainsworth, 1985). Attachment theory proposes that this organization is based
upon real experiences, rather than intrapsychic forces (Bowlby, 1988). The behavioral
organization of interpersonal relationships is fairly stable over time (Sroufe & Waters, 1977).

The child’s relationship behavior consists of two separate, but interconnected systems,
attachment behaviors and exploratory behaviors (Ainsworth, 1964; Main & Solomon, 1986).
When the child perceives a need for closer proximity to the mother, attachment behaviors are
activated. Once the child regains a sense of security, the child is again free to engage in
exploratory behaviors. An harmonious balance between attachment behaviors and exploratory
behaviors is the essence of a secure attachment. An insecure attachment occurs when one set
of behaviors is constantly activated to the exclusion of the other, or when neither attachment

nor exploratory behaviors are successfirlly engaged.

14

The secure attachment pattern is a precursor to close interpersonal relationships and
satisfactory school adjustment, as well as a component of emotional health (Sroufe, 1982).
Infants whose mothers respond sensitively to their crying and other attachment behaviors are
more likely to develop secure attachments than infants whose mothers are less responsive
(Stayton & Ainsworth, 1973). The mothers of securely attached infants hold their infants more
frequently and warmly than the mothers of other infants (Ainsworth, 1979; Main & Stadtmarr,
1981). During brief separations, securely attached infants are often highly distressed; however,
they are easily comforted by their mothers upon reunion (Ainsworth, 1979; Stayton &
Ainswortlr, 1973 ), and are able to use contact with their mothers for “emotional refueling”
(Mahler, 1979).

Developmentally, children who are securely attached in infancy are more cooperative
and less aggressive as one-year—olds (Ainsworth, 1979). In preschool, they are more competent
and empathic in peer interactions, more curious and persistent in problem solving, and better
able to elicit and accept help from others (Ainsworth, 1979; Sroufe, 1982; Sroufe & Waters,
1977). This pattern of interpersonal competence is found in securely attached adolescents and
adults and is associated with various measures of mental health and well being (Kobak &
Sceery, 1988; Rice, 1990).

Insecurely attached infants have mothers who ignore, reject, or otherwise respond
inadequately to their needs, particularly crying (Ainsworth, 1979; Stayton & Ainsworth, 1973).
These mothers often reject or avoid physical contact with their infants (Ainsworth, 1979;

Tracy & Ainsworth, 1981). Insecurely attached infants develop defensive strategies which

15

allow them to maintain proximity to their mother, either by suppressing attachment behaviors
or suppressing exploratory behaviors (Main & Solomon, 1986).

Insecure attachments are associated with a multitude of subsequent social, emotional
and academic dificulties. Insecurely attached infants develop aggressive non-compliant
behavior which continues into childhood (Ainsworth, 1979; Biringen, 1994). Preschool
children previously identified as insecurely attached in infancy Show significant behavioral
aberrations, poor coping and problem-solving skills, and difficulties interacting with peers and
adults (Sroufe, 1982). Insecurely attached first grade boys are less well-liked and have more
behavioral problems than securely attached boys (Cohn, 1990). VVIthin this group, infants who
tend to avoid maternal interaction are subsequently found to lack empathy, Show more
negative affect and aggressive antisocial behavior and engage in hostile or mean interactions
with peers (Sroufe, 1982). A similar longitudinal study finds that children who avoid close
interpersonal relationships at 49 months old have the most severe aggressive and oppositional
behavior problems in elementary school (Crowell, O'Connor, Wollmers, Spraflcin, & Rao,
1991).

The attachment patterns which develop in infancy continue to have a profound impact
upon the quality of subsequent interpersonal relationships throughout the lifespan (Ainsworth,
1989). They become reenacted with others even in the absence of the original attachment
figure. The maintenance of these patterns is dependent upon the internalization of a mental
image of the original matemal relationship (Ainsworth, 1969). These “internal working
models” of the mother and self are based upon the earlier experiences of being parented

(Bowlby, 1988). Internal working models are cognitive structures based upon actual

16

experience which exert a persistent, unconscious influence of expectations, fears, and wishes

regarding interpersonal relationships (Bowlby, 1988).

From Working Models to Object Represean

Several attempts have been made recently to integrate attachment theory and object
relations theory, in particular the concepts of internal working models and object
representations (Levine & Tuber, 1993; Main, Kaplan, & Cassidy, 1985). Attachment
behaviors are the external, observable manifestations of cognitive internal working models
(Main & Solomon, 1986) which are similar in structure and function to object representations
(Bowlby, 1988; Levine & Tuber, 1993; Main, Kaplan, & Cassidy, 1985). The development of
whole object representations depends upon the internalization of a maternal image which
reflects consistent, predictable responsiveness of the mother (Homer, 1984; 1991). Without
“good enough mothering” (\Vrnnicott, 1958), object representations cannot be adequately
formed or internalized (Homer, 1984, 1991). This, in turn, results in what Bowlby has referred
to as an “affectionless character” (Bowlby, 1947). This “inner world of self—other
relationships” (Pine, 1990, p. 102) becomes the blueprint for firture relationships.

A lack of adequately internalized object representations interferes with the development
of appropriate interpersonal relationships and may lead to the development of character
disturbances such as the inability to follow rules, lack of capacity to experience guilt, and
indiscriminate fiiendliness with an inordinate craving for affection with no ability to form lasting
relationships (Homer, 1984, 1991). Individuals who have internalized poorly formed object

representations tend to see others as fluid and interchangeable, or malevolent and cold, and to

17

view the ‘Nvorld as lifeless, alien and unpredictable, as stark and static or fluid and fomrless
(Levine & Tuber, 1993, p. 72).” Impaired object representations have been found in many
clinical populations, including adolescents with conduct disorders, (Weber, 1990; Weber,
Meloy & Gacono, 1992); adults with psychopathic tendencies (Gacono, 1988; Gacono &
Meloy, 1991; Meloy, 1988); and individuals with borderline personality disorder (Kwawer,
1979, 1980) and schizophrenic disorders (Homer, 1984).

Both parental and child factors contribute to the internalization of object
representations (Bowlby, 1988; Call, 1984). However, parental factors play the more powerfirl
role in shaping the quality of the child’s object representations (Goldberg, 1988). Events which
lead to parental absence or stress can interfere with a parent’s ability to respond adequately to
the child’s needs (Deutsch & Erickson, 1989), and therefore impact the child’s object
representations. Furthermore, situations which induce fear, fatigue, or illness in the child
(Bowlby, 1988) may make the child more difficult to comfort, and again may impede a parent
from being appropriately responsive. Traumatic events, particularly early separations between
parents and children, actual or psychological, can have devastating effects on the development
of appropriate object representations (Bowlby, 1977a, 1977b; Homer, 1984). Thus, childhood
trauma and parental childrearing attitudes both play significant roles in the child's internalization
of object representations.

An early study of 10- to 14-year-olds compared children raised in an institution during
infancy to those raised in a foster home from infancy (Goldfarb, 1944). The study concluded
that institutionalized children were more emotionally immature and less able to enter into

interpersonal relationships. This difference was attributed to the “absence of a warm,

18

consistent, continuous, day-to-day contact with an adult in the role of a parenting person” in
the lives of institutionalized children (Goldfarb, 1944, p.446).

Pierce (1978) studied 52 children ages 6 to 17 who experienced the loss of a parent
prior to age seven. This study concluded that children who experienced this loss (due mostly
to divorce or military service) prior to five years of age were more likely to be highly guarded
and defended against dependency needs. This suggests that these youngsters formed
representations of others as unavailable to meet their needs and therefore tended to be reluctant
to become involved in close interpersonal relationships.

Another significant impact of a child’s object representationsis found in cases of early
maltreatment. A longitudinal study 0 “at-risk children” finds that children who had been
abused or neglected in early childhood had greater negative expectations of relationships
(McCrone, Egeland, Kalkoske, & Carlson, 1994). They have “core models of themselves as
unworthy and of others as unavailable, rejecting or exploitative (McCrone et al., 1994, p.
116).” In a study of 43 twelve-year-olds receiving Aid To Families with Dependent Children,
82% of the maltreated group manifested a disorganized attachment pattern, compared to 19%
of the non-maltreated group (Carlson, Cicchetti, Barnett, & Braunwald, 1989). This suggests
that these children had been unable to form consistent, whole object representations which
resulted in their disorganized, inconsistent approach to others.

Childhood trauma also affects the development of object representations. Since object
representations are based upon the accumulation of real life experiences (Pine, 1990), painful,
negative, or otherwise traumatic experiences in young childhood adversely impact upon this

process (Homer, 1984). A sample of 44 children who had been classified as insecurely

19

attached at 12 months showed that those with disorganized attachment behavior patterns, and
therefore the most impaired object representations, experienced more life stressors including
financial problems, having a household member with a long-term illness, or residing in
unsuitable living conditions. Trauma and abuse both interfere with the capacity to engage in
close relationships. A study of 63 children and adolescents who had experienced physical
abuse, sexual assault, accidents requiring medical attention or parental loss through violence all
showed a similar pattern of anger and inability to accept or seek closeness with others

(Holaday, Annsworth, Swank, & VIncent, 1992).

Delimuencyand Traumatic Life Events

Classifications in delinquency reflect “oficial” delinquents, that is, those who have been
oflicially charged with committing a delinquent act, and “self-report ” delinquents, refening
to those who admit committing a delinquent act but may not have come to the attention of the
legal system. Delinquents, both self-reported and ofiicial, experience more stressfirl life events,
especially in early childhood. A multitude of life events have been associated with increased
likelihood of involvement in delinquent behavior. These include broken homes, marital
disharmony, parental psychopathology (Gove & Crutchfield, 1982), parental illness (Kolvin et
al., 1988), and the presence of criminal parents or siblings (Loeber & Dishion, 1983). Bowlby's
study of young thieves (Bowlby, 1947) indicated that they experienced unusually fi'equent and
prolonged separations from their mothers.

Kolvin et al. (1988) conducted a longitudinal study of 847 children born in 1947. They

examined the impact of social and family stressors on the age of first arrest for delinquent

20

behavior. Young delinquents were likely to be arrested more fi'equently and become more
involved in more serious criminal behavior because they have longer criminal careers. Males
whose first arrest occurred prior to age 15 had significantly more early childhood experiences
of parental marital instability, parental illness, poor physical surroundings, social dependency,
overcrowding, and inadequate mothering, significantly more often than males first arrested
after 15 years old.

Sandler and Block (1979) used the Coddington Childhood Life Events Scale
(Coddington, 1972a, 1972b; Heisel, Ream, Raitz, Rappaport, & Coddington, 1973) to
differentiate maladapted from non-maladapted irmer-city youngsters. Maladapted children
were found to have experienced significantly more life changes in the previous year and
significantly more negative life events than their well-firnctioning peers.

Differences in the number of traumatic life events experienced in early childhood, as
measured by the Coddington scales, reliably classified 82.5% of socialized and undersocialized
delinquents with conduct disorder (Deutsch & Erickson, 1989). Traumatic life events were
more predictive of undersocialized conduct disorder than changes in caregivers or child's

perception of parent-child relationships.

Pargmg Attitudesyand Delinquency

Delinquency has frequently been associated with maladaptive parenting pattems,
including physical abuse, inconsistent discipline, harsh physical punishment and lax enforcement
of rules (DiLalla, Mitchell, Arthur, & Pagliocca, 1988; F eehan, McGee, Stanton, & Silva,

1991; Gove & Crutchfield, 1982; Lewis, Lovely, Yeager, & Femina, 1989; Lewis, Shanock,

21

Pincus, & Glaser, 1980). It has also been associated with negative parental attitudes toward
childrearing such as parental rejection of the child (Gove & Cnrtchfield, 1982).

Although delinquency has often been associated with child maltreatment, few studies
actually examine the parenting attitudes of parents with delinquent children. Much of what
appears to be known and accepted as common knowledge about delinquents’ parents is
derived fiom youths’ reports about their parents. While this may be a valuable source of
information regarding delinquents’ object representations, it remains unclear what light these
studies shed on actual parental attitudes toward childrearing.

Adult antisocial and criminal behavior have been associated with a childhood history of
abuse and neglect. In an examination of 908 court records of abuse and neglect, as well as
subsequent adult arrests, Widom found that males who were abused or neglected were
significantly more likely to be arrested for any non-traffic offense and for any violent ofi‘ense
than non-abused, matched controls (\Vrdom, 1989a, 1989b). In a further study, Widom
concluded that childhood abuse or neglect significantly predicted the number of antisocial
personality symptoms in adulthood, even when demographic variables and criminal history
were controlled (Luntz & \Vrdom, 1994).

F arrington (1978) followed 411 males from the age of 8 until 22 years old. This study
found that aggressiveness at ages 8 - 10 was predictive of violent delinquency in young
adulthood. Harsh parental discipline at age 8 was the strongest predictive factor of subsequent
violent delinquency, followed by separations from parents and presence of a criminally involved
parent. In fact, 61.5% of the violent delinquents had experienced harsh parenting compared to

only 27.1% of the nondelinquents.

22

It is not only extreme physical abuse that is related to delinquency. Lesser degrees of
inconsistent or inadequate parenting contribute to delinquent behavior as well. In a study of
childrearing attitudes of parents of 620 thirteen-year-olds fi'om Chicago (Gove & Crutchfield,
1982), the strongest predictor of delinquency was the parents’ feelings toward and perceptions
of the child. Parents who had more negative feelings toward their children provided them with
less supervision, and were more likely to have delinquent children. Harsh physical punishment
and the presence of a single parent family were also more predictive of delinquency, but not as
strongly as negative parental attitudes. This implies that it is not solely the physical pain or
violence associated with abuse that contributes to delinquency. The negative parental attitudes
conveyed with abusive behavior may, in fact, cause the most damage.

This proposition that negative parental attitudes may impact as heavily as harsh,
punitive or abusive behavior appears to be supported by a longitudinal study of 849 mothers
and their children fiom age 7 until 15 years old (Feehan, McGee, Stanton, & Silva, 1991).
Mothers in this study were rated on two dimensions of parenting: strict versus lax, and
consistent versus inconsistent. Counter to the author’s original hypothesis, strictness was not
significantly related to subsequent externalizing behaviors, but inconsistent parenting was. This
is consistent with what Furman and Furman (1984) label “intermittent decathexis”, a
childrearing pattern in which the parent alternates between intense focus on the child, usually
on his or her misbehavior, and sudden, unpredictable disinterest in the child to the point of
being unaware of the child’s whereabouts or activities. Children who experience this type of

parenting are intensely angry, anxious and mistrustful.

23

Several studies have examined the contribution of multiple risk factors to delinquent
behavior. Renkin, Egeland, Marvinney, Mangelsdorf, and Sroufe (1989) have proposed three
correlates of aggressive antisocial behavior: 1) hostile, abusive, or punitive parenting, 2) high
life stress, and 3) history of avoidant attachment in infancy. These factors are theorized to
“interact in such a way as to distort the child’s normal tendency to seek emotional closeness
with others during times of stress or arousal (Renken et al., 1989).” Their study included 191
third grade children and their mothers. The children were part of an ongoing longitudinal study
of high-risk children. They concluded that avoidant attachment history, harsh parental
treatment, and stressfiIl life circumstances predicted aggression in boys, accounting for a third
of the total variance in aggression.

The most extensive studies of factors contributing to delinquency have been conducted
by Rolf Loeber (Loeber & Dishion, 1983; Loeber & Stouthamer-Loeber, 1986). Not
surprisingly, one of the stronger predictors of adolescent delinquency is an earlier history of
antisocial, aggressive or problematic belmvior in childhood (Loeber & Dishion, 1983). In this
meta-analysis, family management techniques were the most predictive factors of delinquent
behavior, increasing the accuracy rate of predicting delinquency by an average efi‘ect size of .50
(Loeber & Dishion, 1983). In a subsequent meta-analysis of longitudinal studies, lack of
parental supervision, parental rejection and conflicted parent-child relationship were the
Strongest predictors of delinquency (Loeber & Stouthamer-Loeber, 1986). Parents’ marital
relationship and history of criminality were moderate predictors. Weaker, but significant

predictors included parental discipline practices and stressful family factors such as parental

24

absence or illness. In both of these studies, socioeconomic status was found to be a poor
predictor of delinquency when these other factors were involved.

In a further examination of the relationship between socioeconomic status and
childhood conduct problems, Dodge, Pettit, and Bates (1994) followed 513 children fiom
preschool through third grade. Socioeconomic status was predictive of conduct problems at all
ages; the strength of this relationship intensified over time. That is, children in the lowest
socioeconomic group were rated as increasingly more aggressive by teachers and peers as they
got older. The effect of socioeconomic status upon aggression was mediated by eight social
and familial factors. These included: harsh parental discipline, lack of maternal warmth,
aggressive adult role models, maternal endorsement of aggressive values, maternal lack of
social support, unstable peer group, lack of cognitive stimulation, and life stressors such as
divorce, death, or legal problems. These factors accounted for 57% of the efl‘ect of
socioeconomic status on teacher-rated conduct problems, and 50% of the variance in peer-
nominated aggression scores. The authors concluded that maternal childrearing attitudes and
stressfirl life events exerted a strong influence on aggressive behaviors from a very young age.
They noted that these early-onset behavior problems are the ones most likely to persist
throughout life (Dodge et al., 1994).

In a reanalysis of an early study on delinquency, Sampson and Laub (1994) concluded
that harsh parental discipline, lack of supervision, and weak parent-child attachment mediated
the effects of poverty, life stress, and prior antisocial behavior on adolescent delinquency. The
reexamination of 500 delinquent and 500 nondelinquent males from low income Boston

neighborhoods shows that “informal social controls,” consisting of the attachment, supervision,

25

and discipline variables, exerted a large, significant negative effect on delinquency and that 68%
of the efi‘ect of family structural variables and “parent/child disposition” (a composite of
parental deviance, instability and child antisocial behavior) on delinquency was mediated by

informal social controls.

Object Representations of Delinquents

Juvenile delinquents, whose behavior is marked by aggression, hostility, and poor
interpersonal relations, have significantly impaired object representations. An early study of
adolescents who committed homicide found these youngsters expected to be harmed by people
in every social interaction, and their “potential for forming fulfilling interpersonal relationships
seemed to be practically nil (King, 1976, p. 698).” These youngsters depersonalize and fail to
differentiate among people (King, 1976), another indication of undifferentiated internal object
representations. In a more formal study of adolescents who commit homicide, 55 adolescents
who committed homicide were compared to 55 non-violent delinquents (Greco & Cornell,
1992). Those adolescents who committed a homicide in the course of committing another
crime evidenced greater disturbance in object representations than delinquents who committed
homicide in the course of an interpersonal conflict and non-violent delinquents. However, all
three groups showed decreased interest in interpersonal relations compared to norms
established for nondelinquent adolescents (Exner, 1986).

When 71 delinquents and 71 nondelinquents were asked for their earliest memory,
delinquents’ memories reflected self-representations of weakness, failure, isolation, and

victimization (Davidow & Bruhn, 1990). They reported fewer memories involving another

26

person. When memories included another, the object representations contained perceptions of
others as untrustworthy, hurtful, frustrating, and unreliable. The study concluded that for
delinquents “there do not seem to be any internalized object relations from which to draw
support (Davidow & Bruhn, 1990, pp. 610 - 611).”

An extensive study of 100 conduct disordered adolescents indicated a disturbed object
relational world with immature personality organization (Gacono & Meloy, 1994). The
predominance of part-objects indicated that for these adolescents, “the integration of good and
bad split objects into whole, integrated objects has not occurred,” which is both a result of and
perpetuates splitting as a defense (p. 67). Furthermore, conduct disordered males gave at least
one response revealing a primitive mode of relating. The distribution of types of primitive
relating responses was virtually identical to that of the adult antisocial personality disordered

group they sampled.

Delinqueng and Perceptions of Parent-Child Relationships

One indication of internal object representations is how an individual perceives others.
Delinquent youngsters perceive their parents differently than do nondelinquent youngsters.
They report perceiving their parents as inconsistent, afi’ectionless, controlling, and less
emotionally available (Brand et al., 1990; Howard, 1981; Mak, 1990, 1991; Rey & Plapp,
1990; Van Voorhis, Collen, Mathers, & Garner, 1988).

Delinquent youth, both self-reported and oficial, reported feeling less bonded to their
parents than their peers (Mak, 1990, 1991). A comparison ofadolescent male first time

ofl‘enders and persistent offenders showed that persistent ofi‘enders reported their mothers were

27

Significantly less caring (Howard, 1981). Conduct disordered and oppositional disordered
youngsters rated their parents as optimally caring only halfas frequently as a control group
(Rey & Plapp, 1990). Both groups rated their parents as more rejecting and controlling than
other youngsters.

Van Voorhis et al. (1988) asked 152 white, rural high school students between 14 and
19 years old about their history of delinquent behaviors, family quality and family structure.
The family quality dimension was comprised of subscales measuring child abuse, affection,
conflict, supervision and overall home quality. Afl‘ection and supervision were the two
subscales most predictive of delinquency. Overall home quality was a significant predictor of
all categories of delinquency except for violent delinquency. Gender was the strongest
predictor of violent crime, strongly outweighing all other variables. If results for males and
females had been analyzed separately, it is likely that many of the family quality subscales
would have significantly predicted delinquency.

F ifly—five adolescents in a residential school for behavior-disordered youngsters were
administered a parental consistency questionnaire (Brand et al., 1990). Youngsters who
showed signs of more severe emotional maladjustment or immature emotional development
perceived their parents as treating them more inconsistently. These results may reflect actual
parenting patterns, but they may also reflect the part objects and splitting referred to by Gacono
and Meloy (1994).

A study of 793 Australian teenagers provides fiIrther clarification on the effects of
adolescents’ perceptions on delinquent behavior (Mak, 1990, 1991). Self-reported levels of

delinquency decreased as parental bonding and empathy increased. Although this study was

28

not based upon an object relations view of delinquency, the results nonetheless support this
theoretical stance. Impaired object representations, which are affected by parental attachment
and decrease the capacity for empathy (Homer, 1984; Meloy, 1988) are a significant factor in
the causation of delinquent behavior.

Several studies have found that delinquent youngsters perceive their parents to be more
rejecting and punitive and their families as experiencing more conflict compared to
nondelinquents. In a study of 244 adolescents from clinical and non-clinical settings, perceived
parental rejection was significantly correlated with self-reported delinquency, even after
controlling for other family variables (Sirnons, Robertson, & Downs, 1989). Furthermore,
using a LISREL model, the authors concluded “that the predominant causal flow is fi'om
parental rejection to adolescent deviance rather than from deviance to rejection (p. 307).”

Similar findings are reported in a larger study of 942 adolescents (Cemkovich &
Giordano, 1987). This study surveyed adolescents on their perceptions of such parenting
variables as supervision, identity support, caring and trust, and intimate communication. As
expected, self-reported delinquents perceived lower levels of caring and trust, supervision,
support and communication. Even when such variables as gender, ethnicity, and household

structure were controlled, internal family dynamics were significant predictors of delinquency.

Sum—mars
Research indicates that the development of internal object representations can be
impaired by inadequate parenting and traumatic events in early childhood. The research on

delinquency suggests that, compared to nondelinquents, delinquents have experienced more

29

early childhood trauma and more harsh, inconsistent or otherwise suboptimal parenting.
Delinquents, as a group, also have fewer internalized object representations relative to their
non-delinquent peers. The representations delinquents do have reflect a view of others as cold,
distant and unresponsive. This can be seen in the differences between delinquents’ and
nondelinquents’ perceptions of parent-child relationships. This study proposed that differences
in object representations between delinquent and nondelinquent adolescent males could be
attributed to differences in early childhood trauma and maternal childrearing attitudes.

F urtherrnore, it proposed that maternal childrearing attitudes would moderate the effect of

childhood trauma on object representations.

CHAPTER 3

Methodology

This chapter describes the participants and measures employed in this study, and
discusses the methodology used for test administration and data collection. In addition, this
chapter describes the research hypotheses and statistical tests that were part of this

investigation.

Participants

The study included 60 boys, ages 13-17, and their biological mothers. Among the
boys, the sample consisted of equal numbers of delinquent and nondelinquent adolescents. The
delinquent sample consisted of 30 boys detained in a secure detention facility. The detention
center is a county-operated facility located in a moderate sized midwestem community. Youth
are detained in this facility on petitions stemming from acts that would be considered felonies if
committed by an adult, thus constituting “serious delinquency.” Participants for the delinquent
sample were recruited during two weekly visitation times at the detention facility.

The nondelinquent comparison group consisted of 30 boys who resided in the same
school district that serves the detention center. Comparison group boys and their biological

mothers were recnrited through local high schools, middle schools and community recreational

30

31

programs. Comparison group boys needed to have passing grades and no history of official
delinquency or disciplinary referrals to the school district’s Ofice of Student Services which
handles serious or repeated infractions of school policy.

An attempt was made to match boys from the two groups with regard to age, race,
receptive vocabulary, and socioeconomic status. Socioeconomic status was measured by
mother’s education level and the boy’s participation in the federally subsidized school lunch
program.

Boys were excluded from the study if any of the following criteria were met: 1) their
receptive vocabulary scores were less than 70 or greater than 130, 2) they sufi‘ered from
psychosis or organic impairment, as assessed by the detention facility psychologist or school
omcials, 3) they were facing waiver to the adult court system, or 4) they had no contact with
their biological mothers. The first two exclusionary criteria were established because
diminished as well as exceptional verbal intelligence and severe psychopathology could have
spurious effects upon the psychometric measures used in this study (Exner, 1986). Those
facing waiver proceedings were excluded because some of the same measures employed in this
study are used in court-ordered psychological evaluations for waiver proceedings and,
therefore, might have been subject to subpoena or otherwise interfered with court proceedings.

Boys who had no contact with their biological mothers were excluded because this study
requires biological mothers to complete measures of early life events and maternal child-rearing

attitudes.

32

Procedure

Mothers of delinquent boys were approached by the investigator before regularly
scheduled visitation hours at the detention facility and asked to participate and to allow their
sons to participate in a study on “factors that affect teenage behavior.” They were informed
that participation was voluntary and would not affect in any way any pending legal proceedings
or the treatment their sons receive while in detention. Mothers were informed that
participation in the study would require about 30 minutes of their time and about 90 minutes of
their sons’ time, and that all responses would be treated confidentially. Upon agreeing to
participate, mothers were asked to sign an informed consent form reiterating the above
information. Mothers were given the measures after the conclusion of visitation. In this way,
the length of visitation was neither prolonged nor curtailed as a result of a mother’s willingness
or refusal to participate.

Boys and their biological mothers were recruited for the comparison group with the
assistance of school principals and local recreation club leaders in accordance with local school
district and recreation club policies. Mothers in the comparison group were provided with the
same information as mothers of delinquent boys. They were also asked to Sign and return an
informed consent form. Upon receipt of the consent form, this investigator arranged to
administer the measures to mothers at a mutually convenient time and location. Arrangements
included home visits, meeting at the school or at the recreation center.

Boys whose mothers consented to participate were asked to participate in the study by
an advanced doctoral student in clinical psychology who administered the measures to them.

The investigation was described as a study on “things that affect how teenagers act” which

33

would take about 90 minutes to complete. Boys were verbally informed that participation was
voluntary, that they could stop at any time, and that their responses would be kept confidential.
Delinquent boys were told that their participation or refusal to participate would not influence
their legal proceedings, length of detention, or level of privileges/restrictions during their
detention. Because of their status as incarcerated minors, particular care was taken to insure
they understood that participation was voluntary and that their consent could be revoked.
Comparison group boys were informed that participation would not afi’ect their grades in
school or their participation in recreation club activities. After verbally underscoring the terms
of participation, boys in both groups were asked to verbally assent to participate. Copies of the
parental consent form and youth assent form can be found in Appendix A
After informed consent was obtained, this investigator administered to mothers the
Children’s Life Events Scale - Preschool version (Coddington, 1972a, 1972b, 1984), the
Adult-Adolescent Parenting Inventory, Form B (Bavolek, 1984) and a demographic survey
designed for this study. Boys were administered the Peabody Picture Vocabulary Test -
Revised, Form L (Dunn & Dunn, 1981), Inventory of Parent-Peer Attachment (Armsden &
Greenberg, 1987) and Rorschach Inkblot Test (Rorschach, 1942) by an advanced doctoral
student in clinical psychology. Measures were given in the same order to all participants.
Most boys were tested during the course of the school day, necessitating that they miss

approximately 60-90 minutes of class time. This may have served as an unintentional incentive
for participation, but was also the most practical means for administering the measures. Boys

who declined to participate returned to their scheduled classes.

34

To avoid examiner bias in eliciting desired responses to the Rorschach test, testing of
all boys was conducted by an advanced doctoral student in clinical psychology who had
completed coursework on the administration and scoring of the Rorschach Inkblot Test using
the Comprehensive Scoring System (Exner, 1986) and who desired additional experience
administering the Rorschach. The examiner was uninformed as to the hypotheses under
consideration, but was not blind as to group membership. The Peabody Picture Vocabulary
Test, Children’s Life Events Scale, Adult-Adolescent Parenting Inventory, and Inventory of
Parent and Peer Attachment were scored by this investigator.

Although this study relied, in part, upon mothers’ retrospective accounts of events
occurring during their youngster’s early childhood, it was believed that mothers would be able
to respond more accurately than the youngsters themselves. This study also assumed that
maternal child-rearing attitudes would have remained fairly consistent over time. It further
assumed that these attitudes preceded, rather than resulted fiom, their sons’ delinquent
involvement. This assumption is supported by the cross-lagged studies which report that
parental attitudes precede involvement in delinquency (Sirnons, Robertson, & Downs, 1989).

Since participants could not be randomly assigned to delinquent or comparison groups,
this study used a quasi-experimental design (Cook & Campbell, 1979). An attempt was made
to match participants as closely as possible on those variables which might affect the results:
age, race, socioeconomic status and verbal intelligence.

Results were kept confidential by assigning a code number to subjects and removing
subjects’ names fiom all measures. All data were stored in a locked file. Group information

and statistics are reported in aggregated form so that individual responses cannot be identified.

35

Mm
Mothers completed a brief demographic form, which can be found in Appendix B, as
well as the Children’s Life Events Scale - Preschool Version (Coddington, 1972a, 1972b,) and
the Adult-Adolescent Parenting Inventory (Bavolek, 1984). Boys completed the Peabody
Picture Vocabulary Test - Revised (Dtmn & Dunn, 1981), Inventory of Parent-Peer
Attachment (Armsden & Greenberg, 1987) and the Rorschach Inkblot Test (Rorschach, 1942).
The measures, which were administered in the above order to each participant, will now be

described more firlly.

Children’s Life Events Scale

The Children’s Life Events Scales (Coddington, 1972a; 1972b; 1984) are a series of
scales designed to identify life events requiring the individual to readjust and adapt to a
substantial change in life circumstances. Separate forms are designed for preschool, elementary
school, junior high school and senior high school populations. The preschool form was used
because this study hypothesized that early childhood events have the greatest impact upon the
development of internal object representations.

The preschool form lists 30 events that may be experienced by children under five
years. The list ranges from fairly common events (i. e., birth of a sibling) to fairly uncommon
events (i. e., death of a parent). Events have a preassigned rating on a scale of 0-100 based
upon the amount of stress they are estimated to cause. The stress estimations were developed
by surveying pediatricians, mental health workers and teachers as to how much adaptation

would be required for a child to cope successfully with a given event, with birth of a sibling

36

being used as a standard (Coddington, 1972a). Rank order correlations of interrater agreement
ranged fi'om .85-.97 (Coddington, 1972a). The interrater correlation between mental health
professionals with 10-19 years direct child care experience and those with over 20 years
experience was .94 (Coddington, 1972a).

The Childhood Life Events Scales were originally developed to detennine whether
medically ill children experienced more stressful events prior to their illness than healthy
children. Subsequent studies (Heisel, Ream, Raitz, Rappapport, & Coddington, 1973) found
that children with physical health problems and children with mental health problems both
experienced significantly more stressfirl life events in the twelve months preceding their
difficulties than did their healthy peers.

Deutsch and Erickson (1989) used the Children’s Life Events Scale with delinquent
youngsters. Undersocialized conduct disordered youth scored significantly higher than
socialized conduct disordered youth, experiencing both more fi'equent and more traumatic
stressfirl life events. No study thus far has compared delinquent youth to nondelinquent youth
on this measure.

The scale was scored according to: 1) the sum total of events experienced, referred to
as “fi'equency” of stressful life events, and 2) the standardized weighted stress value of the
events experienced (Coddington, 19723), referred to as “severity” of stressful events. A copy

of this scale can be found in Appendix C.

37

Agult-Atdolescent Parentmg' Inventory (AAPD

The Adult-Adolescent Parenting Inventory (AAPI; Bavolelg 1984) measures
differences in child-rearing attitudes and beliefs between abusive and non-abusive parents. The
32 items are scored on a five-point Likert scale ranging from (1) = Strongly Disagree to (5) =
Strongly Agree. They comprise four subscales associated with parenting: parental
expectations, parental empathy, value of physical punishment, and parent-child role reversal.

The Parental Expectations subscale measures the parents’ perceptions of the skills and
abilities of their children. It includes such items as, “Children should not be expected to talk
before the age of one year.” High scores indicate a realistic understanding of children and their
developmental limitations, low scores indicate a lack of understanding of appropriate
development. The Empathy subscale measures parents’ awareness of, and willingness to
respond to, the emotional needs of their children. It consists of items such as, “Children will
quit crying faster if they are ignored.” High scores suggest sensitivity to children’s needs. Low
scores are often associated with parents who fear spoiling their children. The Physical
Punishment subscale measures parents’ belief in the acceptability of corporal punishment and
consists of items such as, “Children should always be spanked when they misbehave.” High
scores indicate a disapproval of physical punishment and advocacy of alternative means of
discipline. Low scores indicate an endorsement of physical punishment for responding to
children’s misbehavior. The Role Reversal subscale measures the clarity of parent-child roles
within the family. It includes such items as, “Young children should try to make their parent’s
life more pleasurable.” High scores indicate clear role distinctions between parents and

children with parents assuming the caregiver role. Low scores indicate more diflirse parent-

38

child roles with the child having significant caretaking responsibilities including assuring the
happiness and comfort of parents (Bavolek, 1984).

Separate norms for each subscale are available for abusive and non-abusive parents and
adolescents. Norms are provided for males and females of several racial groups. An
individual’s score can be compared to abusive and non-abusive individuals of the same gender
and ethnicity.

Bavolek (1984) established construct validity through item-construct correlations
(range = .53 - .75), inter-item correlation (range = .17 - .55), and factor analysis (range = .45 -
.68). Internal consistency reliability coeficients were reported to range fiom .75 (Parental
Expectations) to .86 (Physical Punishment) (Bavolek, 1984). Total test-retest reliability for all
items is .76 (range = .39 - .89) over a one-week time span (Bavolek, 1984).

Reliability of the AAPI was established with the sample in this investigation for the
entire scale of 32 items (Cronbach alpha = .85). Reliability of AAPI subscales for this sample
were .44 (Parental Expectations), .74 (Empathy), .64 (Physical Punishment), and .80 (Role
Reversal). Scores on all items were summed for each mother to generate a single “childrearing

attitude” score.

Rorschach flblot Test

 

The Exner Comprehensive System for scoring and interpreting the Rorschach inkblots
(Exner, 1986) is a highly regarded instrument for assessing internalized object representations
(Blatt & Lerner, 1983; Blatt, Tuber, & Auer’oach, 1990; Fritsch & Holrnstrom, 1990; Kwawer,

1979; 1980). A recent review by Fishler, Sperling, and Carr (1990) concluded that the

39

Rorschach Inkblot Test is the most reliable indicator of internalized object representations. It
may be especially useful with delinquent youngsters who typically use massive defensive
operations to maintain positive images of caregivers (W Illoclg 1986) and are therefore unlikely
to respond candidly to self-report measures of interpersonal relationships or parent-child
relations.

Delinquent youth often have impaired object representations and do not readily obtain
gratification fiom close interpersonal relationships compared to nondelinquents. They would
be expected to give fewer responses containing human figures, human movement or texture.

The human content, human movement and texture determinants are among the most
well established of all the determinants in Exner’s scoring system. Test-retest reliability
coefiicients for a non-clinical sample of eight-year—olds over a period of eight years were .54
for human movement responses and .43 for texture responses (Exner, Thomas, & Mason,
1985). Correlation coefficients established with this sample between ages twelve and sixteen
reached .70 for human movement and .77 for texture responses, indicating that these
determinants are stable throughout adolescence.

As only the human content, human movement and texture responses to the Rorschach

were used in the present study, these responses will be discussed more fully.

Human Content (H) Responsi
These are responses whose content includes a pure human form (Exner, 1986).

Responses involving pure human content provide an indication of the individuals interest in

40

people and potential for social relations. The absence of human content “appears to signal a
marked lack of interest in, and/or detachment from people (Exner, 1986, p. 401).”

The lack of at least one pure human content response is virtually non-existent above
age twelve (Exner, 1990). Profiles containing no pure human content responses are extremely
rare among adolescents, occurring in only 0—1% of the records of 12- to 16-year-olds, but
appearing in 16% of conduct-disordered male adolescents (Gacono & Meloy, 1994) and 13%
of the records of character disordered (largely anti-social and borderline personality disordered)
adults (Exner, 1990). The average number of human content responses for a non-clinical
sample of 13- to 16-year-olds ranges from 3.00-3.42 (SD. = 1.71-1.96), however, for conduct
disordered adolescents, the average falls to 2.32 (SD. = 1.95) (Gacono & Meloy, 1994).

Klopfer, Ainsworth, Klopfer and Holt (1954) related human content responses to a
capacity for empathic relationships with other human beings. Human content responses reflect
intemalized images of others which “bear the imprint of [the individual’s] formative
interpersonal history, and reveal something of his ingrained relationship predisposition
(Mayman, 1967, p. 18).” A lack of human content responses is indicative of “some severe
disturbance of [the individual’s] relationship to internalized good objects (Mayman, 1967, p.

19).”

Human Movement (M) Respon_sgs_.
Human movement responses reflect cognitive functions including social skills and

interpersonal interests (Exner, 1986). Mayrnan (1967) has stated that the number of M

responses “may be an index of the subject’s capacity to form empathic interpersonal

41

relationships (p. 21).” Decreased M has been associated with an absence of empathic
responsiveness (Stark, 1966). In a non-clinical sample of 13- to 16-year-olds, the mean
number of M responses ranged from 4.06-4.35 (SD. = 2.13-2.24) (Exner, 1990). For conduct
disordered adolescents, the mean was only 3.30 (SD. = 2.43) (Gacono & Meloy, 1994).

Blatt and Lerner (1983) suggested a relationship between human movement responses
and the quality of internal object representations. Youngsters with few M responses had
poorer relationships with their mothers than those with a high frequency of M responses (Bene,
197 5). Delinquent youngsters would be expected to have fewer M responses than non-

delinquent youngsters.

Texture (T . TF. FD Respongs,

Exner (1986) interprets T as an indicator of dependency needs. The absence of at least
one texture response is a rare occurrence, appearing in only 18% of the records of 13 year
olds, decreasing to 6% of the records of 16 year olds (Exner, 1990). It is found to occur,
however, in 87% of the records of conduct disordered adolescents (Gacono & Meloy, 1994).
The mean number of texture responses for adolescents ranges from 97-106 (SD. = 48-52)
(Exner, 1990). The mean for a sample of conduct disordered adolescents was 0.17 (SD. =
0.45) (Gacono & Meloy, 1994).

Klopfer et al.(1954) suggested that T responses represent an individual’s basic
expectation of affection to be received fiom the outside world. A lack of any T responses

(often referred to as a “T-less protocol”) has been associated with marked deprivation in early

42

relationships, impaired capacity for afl‘ectional relationships and denial or repression of the need
for affection (Exner, 1986; Gacono & Meloy, 1991; Meloy, 1988).

According to Wmnicott (1958, 1965), the infant initially experiences the world through
the sensation of being held. Physical contact becomes the primary means for acquiring
information during infancy and, therefore, contributes to the very earliest notions of self and
others. This view is echoed by others who note that the earliest personal images of others,
especially mothers, are tactile experiences and that texture responses may be thought of as a
representation of these infantile sensory experiences (Maymarr, 1967; Meloy, 1992).

Compared to nondelinquent youth, delinquents would be expected to have a higher
fiequency of so—called “T-less protocols.” Because texture responses violate assumptions of
normal distribution and because of the clinical distinction between records with no T responses
and those with one or more T responses, the texture response variable is treated as a
categorical variable (0 T responses vs. 1 or more T responses) and analyzed through

appropriate non-parametric statistical tests described in the Armlysis section.

Object Representations and the Rorsch_ac_h.

 

Several studies have shown that youngsters who have experienced inadequate
parenting have fewer human content, human movement and texture responses (Brecher, 1956;
Goldfarb, 1944; Holaday, Annsworth, Swank, & Vmcent, 1992; Monalto, 1952; Pierce,
1978). Youngsters whose parents were psychologically unavailable Show similar patterns of

disinterest in interpersonal relationships and denial of dependency needs.

43

Children ages six and seven whose parents were most restrictive and rule-oriented had
fewer human movement and human content responses than those whose parents were less
restrictive (Monalto, 1952). Those whose parents were restrictive and cold also had fewer
texture responses compared to those whose parents were restrictive yet warm. The results of
this study suggest that those children who experienced unafl‘ectionate, restrictive parenting had
decreased interest in interpersonal relationships and needed to defend against dependency
needs because they lacked well developed internal object representations.

Brecher (1956) obtained similar results when comparing 25 maternally overprotected
and 25 maternally rejected adult male schizophrenics. Those experiencing maternal rejection
were interpersonally distant. They avoided close, intimate relationships. Matemally rejected
subjects gave fewer human movement and texture responses than those who were
overprotected, indicative of impaired internalized object representations.

Research with the Rorschach has shown that it is important to control for the number
of responses per protocol (Exner, 1991). The number of human content responses and the
number of human movement responses were calculated as a percentage of all responses to
control for the effects of response productivity (Exner, VIglione, & Gillespie, 1984). A single
“internalized object representations” score was calculated as the combined percentage of
human movement and human content responses (Blatt & Bennan, 1984; Blatt, Brennis,
Schimek, & Glick, 1976; Exner, Vrglione, & Gillespie, 1984; Fritsch & Holstrom, 1990).
Several published studies have combined multiple Rorschach determinants to construct single
scores for complex constructs, particularly in the assessment of object representations and

object relations (Athey, Fleischer, & Coyne, 1980; Fritsch & Holstrom, 1990; Spear, 1980).

44

Frequency of texture responses was not included due to the non-normal distribution of this
response in the general population. Delinquents were expected to score lower on this measure

of internalized object representations than nondelinquent boys.

Delinquent; and the Rorschach

Studies that use the Rorschach test with delinquents have found Significant differences
in the number of human content, human movement and texture responses between delinquent
and non-delinquent adolescents. A study of 30 Japanese delinquent males (DeVos, 1979)
found that delinquents gave fewer human content, human movement and texture responses
than a control group of 20 nondelinquent males. The mothers of delinquent boys also gave
significantly fewer texture responses compared to mothers of nondelinquents. Mothers and
fathers of delinquent boys had Significantly fewer human movement responses, compared to the
parents of non-delinquent males.

Schachtel (1951) reported that 59.2% of 500 delinquents gave no M responses,
compared to 51.8% of non-delinquent controls. Decreased human movement responses have
been reported for delinquents in India as well (Majumber & Roy, 1962; Ray, 1963;
Raychaudhuri & Maitra, 1965).

Fewer human content responses have been reported among various types of
delinquents. Both convicted and “incipient” (non-adjudicated) delinquent thieves had lower
human content (Raychaudhuri & Maitra, 1965). Non-violent delinquents under ten years old
gave as few M and H responses as older, serious ofl'enders (Majumber & Roy, 1962),

suggesting pre-adolescent involvement in delinquency may be indicative of more severe

45

developmental breakdown than later involvement in delinquency (Steinberg, 1986). Sinha,
Singh, and Singh (1986) also reported finding fewer H responses among a heterogeneous
group of delinquents in India. In a sample of youngsters charged with status offenses (offenses
which are illegal due to the individual’s status as a minor, and which are not as likely to be
adjudicated today, i.e., truancy, incorrigibility, running away), Endacott (1941) found
decreased M, but nomral H. The percentage of human content responses was found to relate
to both severity of crime and recidivism (Curtiss, F eczko, & Marohn, 1979).

To summarize expected findings on the Rorschach Inkblot Test, delinquent boys were
expected to have fewer human content and human movement responses than nondelinquent
boys, as delinquent behaviors are indicative of non-empathic social relationships with others.
Delinquent boys were expected to have a higher frequency of “T-less” protocols than
nondelinquent boys. Human content and human movement responses were also expected to
relate to the frequency and severity of stressful life events and to maternal childrearing

attitudes.

Invenyogy of Parent and Peer Attachment

The Inventory of Parent and Peer Attachment (IPPA) (Armsden & Greenberg, 1987)
assesses the degree to which adolescents are attached to parents and peers. The IPPA is a self-
report measure containing 28 items addressing parental attachment and 25 items regarding peer
attachment. Answers are scored on 5-point Likert scale ranging from (1) = “Never or Almost
Never” to (5) = “Always or Almost Always.” The scale is more closely based on Bowlby’s

theory of attachment than many other self-report attachment measures (Lopez & Gover,

46

1993). Attachment in this measure is defined as “an enduring afi'ectional bond of substantial
intensity (Armsden & Greenberg, 1987, p. 428).” The security of this afi‘ectional bond rests in
the continued accessibility and responsiveness of the attachment figure.

The IPPA assesses the cognitive expectancies that form the adolescent’s intemal
working model of attachment figures. The IPPA contains separate sections for assessing
parent and peer attachments; however, this study only used the section pertaining to maternal
attachments. Attachments are measured along three dimensions: trust, communication and
alienation. The Trust subscale reflects understanding and respect. The Communications
subscale assesses the amount of verbal communication and the Alienation subscale measures
anger and isolation (Lopez & Gover, 1993).

Internal reliability for the three parent scales using Cronbach’s alpha range from .91
(parent trust and communication) to .86 (parent alienation) (Armsden & Greenberg, 1987).
The scales were highly intercorrelated, ranging fiom .76 to -.40 (Armsden & Greenberg,
1987). In the present study, internal reliability for this scale using Cronbach’s alpha was .93.
The reliability of individual subscales was not investigated in this study.

Attention disorders, conduct problems and depression in adolescents have been shown
to be significantly related to overall lower perceived parental attachment as measured by the
IPPA (Nada Raja, McGee, & Stanton, 1992). Adolescents reporting lower levels of parental
attachment on the IPPA also reported experiencing more negative life events and more distress
(Nada Raja et al., 1992).

In the present study, youth were instructed to complete the IPPA with respect to their

relationship with their mothers. A summary difference score of the three parent subscales

47

(trust + communication - alienation) was used as a score of “perceived perception of mother-
child relationship.” Although the preferability of using individual subscales is noted (Lopez &
Gover, 1993), the reliability and validity studies conducted with the summary measure indicate
substantial suitability of the summary score for these analyses (Armsden & Greenberg, 1987).
Delinquent youth were expected to report significantly more troubled attachments to
their mothers. As with the Rorschach, the results of the IPPA were expected to be related to
frequency and severity of stressful life events and maternal childrearing attitudes for both

groups of boys. A copy of this measure can be found in Appendix D.

my Picture Vocab_ul_ary Test - Revigd

The Peabody Picture Vocabulary Test - Revised (PPVT-R; Dunn & Dunn, 1981) is a
brief measure of receptive vocabulary which has been highly correlated with verbal intelligence
(Dunn & Dunn, 1981). The PPVT-R is specifically recommended for use “as an initial
screening device” for identifying exceptionally bright or slow individuals (Dunn & Dunn, 1981,
p. 3). The revised edition of the PPVT was standardized on a nationwide sample of 4200
children based on demographic information fiom the 1970 US Census. The sample was
representative of the population with regard to age, race, geographical location, parent
occupation, ethnicity, and community size (Dunn & Dunn, 1981).

Test developers reported that split-half reliability coefficients for youth between the
ages of 12- to 17-years-old, determined by the Rasch-Wright latent trait method, ranged from

.83 to .87. Alternate form test-retest reliability coefficients for this age group ranged fiom .76

48

to .88. Delayed alternate forms retest reliability coeficients ranged from .68 to .88 for a time
period of9 to 31 days (Dunn & Dunn, 1981).

The median correlation between the PPVT and numerous vocabulary tests, including
vocabulary subtests of individual intelligence tests, was .71 with a range from .20 to .89 (Dunn
& Dunn, 1981). The PPVT had a median correlation with the Stanford-Binet Vocabulary
subtest of .72 (Dunn & Dunn, 1981). Correlations with the Wechsler Intelligence Scale for
Children (WISC) vocabulary subtest ranged fiom .37 to .83, with a median correlation of .69
(Dunn & Dunn, 1981). The PPVT has also been correlated with scores on individual
intelligence tests which ranged from .38 on a non-verbal test to .72 on the Full Scale of the
WAIS. The median correlations of the PPVT with Verbal and Full Scale scores on the WISC
were .66 and .64, respectively (Dunn & Dunn, 1981). Due to restandardization of scores on
the PPVT-R, the authors suggested that differences in standard scores between the PPVT-R
and WISC-R “should be more than eliminated (Dunn & Dunn, 1981, p. 63).”

In a sample of delinquent youth, the correlation between the PPVT and the WISC was
found to be .70 for the Verbal Scale, .69 for the Performance Scale and .73 for the Full Scale
(Kendall & Little, 197 7). Slightly higher correlations were found for an adolescent psychiatric
sample (DeHorn & Klinge, 1978). The correlations between the PPVT-R and WISC-R for
this sample were reported as .79 for Verbal Scale, .65 for Performance Scale and .7 8 for Full
Scale. Although both studies conclude that the PPVT-R is an inadequate substitute for a
complete intellectual assessment battery, the authors agree on its utility as a screening device,

especially at lower levels of functioning (DeHom & Klinge, 1978; Kendall & Little, 1977).

49

This measure was used to screen out youngsters with unusually high or low receptive

vocabulary and to insure that groups were matched with regard to verbal intelligence.

Remh Hypotheses

 

The following hypotheses were investigated in the present study:

1. H1:
2 H2:
3 H3:
4 H4:
5 H5:

Delinquent youngsters have experienced a greater number of stressful life
events than nondelinquents as measured by the number of stressful events on

the Children’s Life Events Scale - Preschool version.

Delinquents have experienced more severe traumatic life events than
nondelinquents as measured by the weighted score on the Children’s Life

Events Scale - Preschool version.

Mothers of delinquent youngsters will score lower (the more abusive direction)
than mothers of nondelinquents on the Adult-Adolescent Parenting Inventory,

Form B.

The frequency of no texture responses on the Rorschach Inkblot Test will be

greater for delinquent youth than nondelinquent youth.

Relative to their nondelinquent peers, delinquent youngsters will report

perceiving fewer human figures on the Rorschach Inkblot Test.

10.

H72

H92

H102

50

Relative to their nondelinquent peers, delinquent youngsters will report
perceiving less human movement than nondelinquent youngsters on the

Rorschach Inkblot Test.

Internalized object representations can be predicted fiom both early childhood I
life events and maternal childrearing attitudes as measured by human content
and human movement responses to the Rorschach Inkblot Test, Children’s Life
Events Scale - Preschool version, and Adult Adolescent Parenting Inventory,

Form B.

Maternal attitudes toward child rearing, as measured by the AAPI, F orrn B,
will moderate the effect of traumatic life events on the internalized object
representations of youngsters as measured by human content and human

movement responses to the Rorschach Inkblot Test.

Compared to nondelinquent peers, delinquent youngSters will perceive
themselves as less attached to their mothers as measured by the Inventory of

Parent - Peer Attachment.

Perceived quality of mother-chfld attachment relationship, as measured by the
Inventory of Parent and Peer Attachment - Parent Scale, will be predicted from

both early childhood life events and maternal childrearing attitudes, as

51

measured by the Children’s Life Events Scale - Preschool version and the Adult

Adolescent Parenting Inventory, Form B.

11. H”: Maternal attitudes toward childrearing, as measured by the AAPI, Form B, will
moderate the effect of traumatic life events on youths’ perceived mother-child
relationship, as measured by the Inventory of Parent and Peer Attachment -

Parent Scales.

Am

Difl‘erences in the frequencies of texture responses on the Rorschach Inkblot Test were
determined by chi square tests for two independent samples. Despite criticisms that non-
parametric tests decrease the statistical power and are over-used in Rorschach investigations
(Acklin, McDowell, & Omdofl‘, 1991), a non-parametric test of categorical variables is
required because of the important clinical difi’erences between profiles with no texture
responses and those with one or more responses (Weber, 1990). A 2 x 2 contingency table,
with zero and one or more responses as categories, was used to structure the data.

A one-way Multivariate Analysis of Covariance (MAN COVA) , using maternal
educational level as a covariate, was conducted to determine between group difi‘erences
(delinquent versus nondelinquent) on the following dependent variables: 1) stressful life events
- frequency, 2) stressful life events - severity, 3) maternal childrearing attitudes, 4) percentage
of human content, 5) percentage of human movement, and 6) perceived mother-child

relationship. Univariate Analyses of Covariance (ANCOVAS) were used to test the

52

significance of each dependent variable individually. The advantage of using a single
MANCOVA over multiple AN COVAs to control Type I and Type H errors has been well
established, especially if the dependent variables may be highly intercorrelated (Bray &
Maxwell, 1982; Haase & Ellis, 1987).

Hierarchical multiple regression (W ampold & Freund, 1987) was used to test the
hypotheses that the effect of life events (fiequency and severity) on internal object
representations is moderated by maternal parenting attitudes and that this model is a significant
predictor of both object representations (combined percentage of Rorschach H and M
responses) and perceived mother-child relationship as measured by the composite IPPA score.

All tests were conducted at an alpha level of .05 and were performed using the most

recent version of the Statistical Package for the Social Sciences (SPSS—X) computer software

program.

CHAPTER4

Results

This chapter presents the results of this study in two parts. The first part consists of an
examination of demographic variables for the delinquent and nondelinquent groups. The

second section presents the results of the hypothesis tests.

Demographic Variables

The delinquent group consisted of 9 Caucasian boys and 21 minority boys. Their mean
age was 15.9 years old (SD. = .91). Twenty of the boys participated in the federally
subsidized school lunch program, 10 did not. Their average score on the Peabody Picture
Vocabulary Test (PPVT-R) was 84.93 (SD. = 16.08) and their mothers completed 11.6 years
of school, on average (SD. = 2.43). Five of the boys in this group lived in an intact family with
both biological parents. Ten lived with only their mother and 15 lived with their mother and
step-father or other adult.

The comparison group of nondelinquent boys consisted of 23 Caucasian boys and 7
minority boys. Their mean age was 14.7 years (SD. = .84). Ten boys participated in the
school lunch program, 20 did not. Their average score on the PPVT-R was 98.1 (SD. =

17.72) and their mothers completed an average of 13.9 years of school (SD. = 2.50). Eleven

53

54

Table 1

Distribution of Demographic Va__ri_ables

 

 

 

Delinquent (n=3 0) Nondelinquent (n=3 0)

Variable N % N %
Age

14 2 ' 6.7 15 50.0

15 7 23.3 10 33.3

16 12 40.0 4 14.3

17 9 30.0 1 3.3
Ethnicig

Caucasian 9 30.0 23 76.7

Afiican American 15 50.0 6 20.0

Hispanic 5 16.7 1 3 .3

Other 1 3 .3 -- --
Lunch Subsidy

None 10 33.3 20 66.7

Partial 1 3.3 2 6.7

Full 19 63.3 8 26.7
Resides with:

Mother only 10 33.3 11 36.7

Mother and Father 5 5.0 11 36.7

Mother and Step-Father 4 4.0 8 26.7

Mother and Other Adult 3 10.0 -- --

Other 8 26.7 -- ~-

 

55

boys came from intact families, 11 lived with only their mother and 8 lived with their mother
and another adult. The distribution of demographic variables can be seen in Table 1.

Compared to the nondelinquent group, the delinquent boys were significantly older
(t(58) = 5.47, p < .001). The delinquent group also had lower receptive vocabulary scores
(t(58) = -3.01, p < .01) and their mothers had fewer years of formal education (t(58) = -3.67, p
< .01). Delinquent boys were more likely to be minority (x21CN=60) = 13.13, p < .001) and to
receive federally subsidized school lunches (121(N=60) = 6.67, p < .01). With regard to whom
boys lived with, there was no significant difference between delinquent and nondelinquent boys
(x2‘(N=60) = 4.43, p > .11).

Because of the small representation of Hispanic and other minority boys in both
groups, non-White boys were combined into as single category labeled “Minority,” with the
remaining White boys referred to as “Majority” on a dichotomous majority/minority status
variable. Similarly, participation in the federal school lunch program was dichotomized into
non-subsidized/subsidized. These dichotomous variables were incorporated into subsequent
analyses with delinquent, majority, and nonsubsidized groups all having values of 1 and
nondelinquent, minority, and subsidized groups having values of 2. The means, standard
deviations and ranges for dependent variables are listed in Table 2.

A correlation matrix indicated that demographic variables were highly intercorrelated
with each other and with many of the predictor variables. Minority status was significantly
correlated with lower levels of maternal education and lower PPVT-R scores as well as
participation in the Federally subsidized school lunch program (all ps < .01). This indicates that

boys from minority groups were likely to be economically disadvantaged, have more limited

Table 2

56

Descriptive Statistics of Dependent Variables (N = 60)

 

 

Variable M SD. Minimum Maximum
Rorschach R 18.22 8.50 5 44
Rorschach H 1.83 1.80 0 8
Rorschach M 2.58 2.09 0 10
Childrearing Attitude 112.12 18.00 41 146
EPA Score 104.08 20.07 51 134

 

57

verbal skills, and have mothers who had less formal schooling. Mothers’ education level was
negatively correlated with lunch subsidy, indicating that boys with more educated mothers
came fiom families which were more economically advantaged.

Trauma fiequency and intensity were highly intercorrelatated (p < .01), which was
expected, as the intensity scale was based upon the weighted sum of the number of traumatic
events. Childrearing attitudes were highly correlated with maternal education and boys’
PPVT-R scores (ps < .01), suggesting a relationship between positive childrearing attitudes and
academic success for mothers and sons. PPVT-R scores were also mildly associated with
ethnic status (p < .05). Human movement and human content scores were highly
intercorrelated with each other and with the number of Rorschach responses (ps< .01 ),
indicating the appropriateness of calculating percent of responses represented by human
movement (M) or human content (H) and using these percentages in subsequent hypothesis
testing. Human content, human movement, and EPA scores were not correlated with
delinquency status or any other predictor variable, a surprising finding.

For the purpose of this study, only 2 variables, mq'ority/minority group status and
mother’s education level, were incorporated into subsequent analyses as the most statistically
and theoretically relevant. Both variables were statistically related to delinquency as well as to
the dependent variables of interest, whereas age, though correlated with delinquency, bore no
theoretical or statistical relationship to the dependent variables. The correlation matrix can be

found in Table 3.

58

 

Evanstnqvmu. ”Bus-Nuausax
.686. gmumg. 2:2. nausea. mes-HMSEEEINanzucésmESE ”ASBEOZINasEonucséeisoz

 

 

I 28m 5% .2
2. I 2385— .2
3. .38. I Egon—om .:
3. :8. :2. I «38.2 .2
8.- 8. 2. a. I snag magic .8
8.- 2. 8. 2. 8. I 58.5 maze-r .w
8.- 2. 8. S. 8. :8. I Susan 5:25 s
8.- cm. 8. 8. :3. 8. 8. I 28m .58 e
8.- 8 - 3 2. n - 2 a :2.- I 82% :83 .m
8. 8. 8. 8. :8. 8.- 8. :8. .2.- I .23 em 9582 .4
z. 8.- 8.- 2. 8n- 8. MN. :9..- :mm. :9..- I 35m €85 .n
2. 2 .- 2.- 2.- 8.- 2. 2. :8.- ..8. :- 8. I a? .N
2.- 3. 8. 8.- ...3“. 1.3.- 1.2.. IBM. :mmr esmv. .3..th eswmr ! EA

2 S : 2 a w a e m a. m N _

 

Asuzvflfi5>a88§=§§é§8§§so
33

001
ma

Mi

4

I101

 

 

 

 

Sta

and

59

Results of Hypothesis Testing

A 2 x 2 (Group x Minority Status) Multiple Analysis of Covariance (MAN COVA) was
conducted using maternal education level as a covariate. The analysis indicated a significant
main effect for Group (WIlks’ 136,50 = 2.61 p < .05), but no significant multivariate effect for
Minority Status or Group by Minority Status interaction. The latter, however, approached
Significance (erks’ E550 = 2.14, p < .07). Two of the seven subsequent univariate
ANCOVAS for the Group main effect indicated significant difi‘erences between delinquent and
nondelinquent boys. A significant univariate ANCOVA was found for the Group by Minority
Status interaction on childrearing attitudes (131,55 = 8.80 p < .01). Means, standard deviations

and F-ratios for dependent variables included in the MANCOVA can be found in Table 4.

Traumatic Life Events.

 

Univariate AN COVAs indicated that delinquent and nondelinquent boys differed
significantly with regard to the frequency and severity of early childhood trauma they
experienced. Delinquent boys experienced both more fiequent and more severe stressfill life
events prior to four years of age.

The first hypothesis concerned the fiequency of traumatic events and stated that
delinquent boys would experience a greater number of traumatic events in early childhood than
nondelinquent boys. This was measured by the Children’s Life Events Scale - Preschool

version (CLES) which was completed by mothers. Delinquent boys experienced significantly

60

Table 4

Analyses of Va_rgnce for Delinquents and Nondelinquents

 

Delinquent (n=3 0)

 

Nondelinquent (n=3 0)

 

 

 

 

Majority Minority Majority Minority

M S_D M S_D M S_D M 5.12 Eratio
Trauma 2.11 0.93 2.71 2.28 1.35 1.19 1.23 1.11 7.24"
Frequency
Trauma 107.78 44.16 114.57 95.33 55.17 44.51 58.71 46.60 8.97“
Intensity
Childrearing 100.78 16.27 108.33 12.58 123.52 10.91 100.57 31.13 8.80““
Attitude
Human 9.23 8.40 8.22 8.30 11.42 8.27 14.89 20.86 1.90
Content
Human 12.09 11.32 12.40 9.95 16.10 11.42 18.54 19.06 2.05
Movement
EPA Score 101.22 21.71 110.24 19.14 101.65 17.32 97.29 28.12 1.41

 

Note. " represents Group x Minority Status interaction ratio

**=p<.01

 

61

more traumatic events in early childhood events than nondelinquent boys (151,55 = 7.24, p <
.01).

The second hypothesis concerned the severity of early childhood trauma and stated that
delinquent boys would experience more severe trauma in early childhood. Weighted values of
the CLES items were used to test this hypothesis. Significant differences were found between
delinquent and nondelinquent boys (151,55 = 8.97, p < .01). Delinquent boys experienced

significantly more severe early childhood trauma than nondelinquent boys.

Maternal Childrearing Attitudes.

Mothers of delinquent boys were expected to have more abusive childrearing attitudes
than mothers of nondelinquent boys. Childrearing attitudes were measured by the Adult-
Adolescent Parenting Inventory (AAPI). The univariate ANCOVA indicated no Significant
differences between delinquent and nondelinquent boys on this measure (151,55 = 0.59, p > .44).

The near-significant multivariate Group by Minority Status interaction (W Ilks’ £6,” = 2.14, p

< .07) was largely due to the significant univariate AN COVA observed for maternal
childrearing attitude (131,55 = 8.80, p < .01). Within the group of mq'ority boys, mothers of
delinquents endorsed harsher childrearing attitudes than mothers of nondelinquents (Muir-4m
= 100.79, SD = 16.28 vs. Munch...“ = 123.58, SD = 10.91). For minority boys, the mothers
of nondelinquents endorsed harsher attitudes than the mothers of delinquent boys
nondelinquents (Mam... = 108.33, SD = 12.58 vs. Murdering-Ian = 100.57, SD = 31.13).

To explore these relationships firrther, a subsequent 2 x 2 (Group x Minority Status)

MANCOVA of the four AAPI subscales was conducted. This additional analysis indicated a

62

 

 

S. v a n 4... .3. v a n .._ ”pong-N u 4.88 =< .802

 

 

 

 

 

:8. :8.- -- 84 2 8.8 8. _ m 3358 208 .4
8. 8.- :8. - 44 2 40.8 8.8 855258
8. :8.- :8. :8. -- 84 8 mos 8.8 €858
8. 2.- :84. :8. :8. - 8 m 8.4 8.8 8868me
sow. seemr eehw. .4er. 15¢. same. 03 :4 05d" Rpm: ofiom~m<<
- 8T5
“possum—ovcoz
:48. 2. -- 8 2 8.8 8.8 8858 Box .4
8.- 2. 8. -- 8 8 8.4 8.8 Bosses:
...8. 8. :8. 8. - 8 8 8.8 8.8 8858
:8. 2. :48. 8. :8. - 8 2 88 8.8 8888me
:8. 8. :8. :8. :8. :8. 8_ 8 8.2 8.82 08282
8T5
along
:8. :8..- -- 84 2 2 .o 8.8 8888 2cm .4
...8. 8.- :8. I 44 2 8.8 8.8 Heofiaaaa
:8. 8.- :8. :8. -- 84 m 8.8 8.8 8888-":
:84. 2.- :8. :84. :8. I 8 m 8.4 8.8 88888:
:8. :8.- :48. :8. :8. :8. e: #4 8.: 2.8: 0.402%
8sz
038mm Pumas—m
834m mesa
Base: 5852 4 m N a as: a: ma 8

 

 

mofiomnzm 588220 88082 «e 205%»8825 v3 SEE-3m gun-Bomom

m Baa-H.

63

significant Group by Minority Status interaction (VVIlks’ 134,52 = 3.15, p < .05) on the omnibus
MANCOVA, but no Significant findings for the main effects of Minority Status (W'llks’ 54,52 =
0.92, p > .46) or Group (VVIlks’ 134,52 = 1.63, p > .19) on this measure. Subsequent univariate
ANCOVAs of the interaction efl’ect were significant for two of the four subscales of the AAPI.
Interaction effects were observed for subscales measuring maternal empathy (131.55: 10.14, p <
.01) and role reversal (131,55 = 8.60, p < .01). For majority boys, delinquency was associated
with decreased maternal empathy and greater role reversal. For minority boys, delinquency
was associated with higher levels of empathy and decreased role reversal. As can be seen fiom
Table 5, Minorin Status was associated with maternal childrearing attitudes for nondelinquents
only. For delinquent boys, maternal education level was substantially more correlated with
childrearing attitudes and the AAPI subscales are less intercorrelated, suggesting less consistent

patterns of responding to test items among mothers of delinquent boys.

Object Representa_tio_n_s_.

Delinquent and nondelinquent boys were expected to difi‘er Significantly on several
Rorschach variables related to object representations. Specifically, delinquent boys were
expected to have a greater frequency of protocols without texture responses and give fewer
human content or human movement responses. No significant difi‘erences were found between
delinquent and nondelinquent boys on the number of human figure responses (131,55 = 1.90, p >
.17) or the number of human movement responses (131,” = 2.05, p > .15) given on the

Rorschach Inkblot Test.

Table 6

64

Summary of Hierarchical Regression Analysis for Variables

Predicting Opiect Representations CN=®

 

 

Variable _B_ SE B [3
Step 1

Group 906.59 653.91 0.22
Step 2

Trauma Frequency 1645.54 939.01 1.34

Trauma Intensity -21.93 40.48 -0.7 5
Step 3

Childrearing Attitudes 17.25 21.20 0.15
Step 4

Interaction 1 -13.07 7.69 -1.17
(Trauma Frequency x
Childrearing Attitude)

Interaction 2 0.16 0.34 0.63
(Trauma Intensity x
Childrearing Attitude)

 

Note. R2 = .04 for Step 1; AR2 = .01 for Step 2; AR2 = .00
for Step 3; AR2 = .05 for Step 4 (all as > .05)

 

65

Delinquent boys were expected to have Rorschach protocols without texture responses
(T = 0) more fiequently than nondelinquent boys. A 2 x 2 contingency table was constructed
with categories of 0 and 1 or more texture responses. Chi square analysis indicated no
significant difference between delinquent and nondelinquent boys in the frequency of “T-less”
protocols (fa-=60) = 0.11, p > .73), disconfirming this hypothesis.

A hierarchical multiple regression analysis was conducted to test the hypotheses that
object representations could be predicted fi'om a combination of maternal childrearing attitudes
and severity and fi'equency of traumatic early childhood events and that maternal childrearing
attitudes moderated the efl‘ect of early childhood trauma on object representations. The
regression analysis shown in Table 6 did not support either hypothesis.

Delinquent and nondelinquent boys were hypothesized to differ with regard to their
perceptions of their attachment relationships to their mothers. The perceived quality of the
mother-son attachment relationship was measured by the Inventory of Parent-Peer Attachment
(EPA). A Univariate AN COVA found no significant difference between delinquent and
nondelinquent boys on this measure (51.55 = 1.41, p > .23). It was also hypothesized that boys’
perception of their attachment relationship with their mothers could be predicted fiom maternal
childrearing attitudes as well as frequency and severity of traumatic life events, and that
maternal childrearing attitudes would moderate the effect of traumatic childhood events on the
adolescent’s perception of the mother-child attachment relationship. This hypothesis was
tested as part of the hierarchical multiple regression shown in Table 7. The results indicated
that, relative to other variables, delinquency was the strongest, and only Significant, predictor of

EPA scores. Maternal childrearing attitudes and early childhood trauma were not found to be

Table 7

66

Summary of Hierarchical Regression Analysis for Variables
Predictig EPA Scores (N=@

 

 

 

Variable l_3 SE 1;
Step 1

Group -12.83 6.28 -0.32*
Step 2

Trauma Frequency 1.00 9.02 0.09

Trauma Intensity -0.60 0.39 -2. 12
Step 3

Childrearing Attitudes -0. 14 0.20 -0. 13
Step 4

Interaction 1 -0.00 0.07 -0.02
(Trauma Frequency x
Childrearing Attitude)

Interaction 2 0.00 0.00 1.90

(Trauma Intensity x
Childrearing Attitude)

 

Note. R2 = .03 for Step 1; AR2 = .02 for Step 2; AR2 = .00
for Step 3; AR2 = .04 for Step 4 (* = p < .05)

 

67

significant predictors of EPA scores, nor was maternal childrearing attitude found to moderate
the effects of early trauma on EPA scores. These two hypotheses were, therefore, not

supported.

CHAPTER 5

Discussion

This chapter summarizes and discusses the finding of this study and offers
suggestions for fiIture research. The first section of this chapter summarizes research
findings of this investigation and relates these findings to existing research. The second
section discusses implications of this investigation for theory and clinical practice. A
discussion of the study’s limitations and suggestions for future research are contained in

the final sections.

Summm

This study used attachment theory as a basis for suggesting that traumatic early
childhood events and harsh, indifferent, or abusive maternal childrearing attitudes
contribute to delinquency by affecting the child’s internal working models or object
representations. Thirty delinquent and thirty nondelinquent males completed a test of
receptive verbal ability, a measure of perceived mother-child relationship, and the
Rorschach Inkblot Test. Their biological mothers completed a demographic information
form, a checklist of traumatic childhood events their sons may have experienced and a

measure of childrearing attitudes. All sixty boys resided in a mid-sized Midwestem city.

68

69

Attempts were made to match the groups on age, race, socioeconomic status, and verbal
ability.

Substantial significant differences existed between the delinquent and
nondelinquent samples on many of the demographic variables. Compared to the
nondelinquent group, the delinquent boys were older, poorer, and were more likely to be
from a minority group. They had lower verbal comprehension and their mothers were less
well educated. They were just as likely as nondelinquent boys to come from intact, step-
parent, or single-parent homes. Race and mother's educational level were used as
covariates in subsequent analyses in an attempt to minimize the effects that differences in
demographic variables would have upon the results of hypothesis tests.

A Multivariate Analysis of Covariance (MAN COVA) indicated that delinquent and
nondelinquent boys differed significantly on the set of key measures used in this study.
Univariate Analyses of Covariance (AN COVAs) indicated that delinquent boys had
experienced more frequent trauma in early childhood and more severe early childhood
trauma than did their nondelinquent counterparts. This finding is consistent with previous
research findings which relate early childhood trauma to aggressive acting out (Loss,
Beck, & Wallace, 1995), unsocialized delinquency (Deutsch & Erickson, 1989; Kolvin,
Miller, Fleeting, & Kolvin, 1988), and impaired emotional and physical health (Heisel,
Ream, Raitz, Rappaport, & Coddington, 1973). Particularly noteworthy was the finding
that trauma was not significantly correlated to either minority status or participation in the

Federally subsidized school lunch program, an indith of socioeconomic status. This

70

suggests that early childhood trauma occurs without regard to race or socioeconomic
status.

The mothers of delinquent boys were expected to endorse harsher, colder attitudes
toward childrearing as measured by the Adult Adolescent Parenting Inventory (AAPI,
Bavolek, 1984). The results of this study suggest that the relation of maternal childrearing
attitudes to delinquency is a complex issue. In the present study, neither delinquency nor
minority group status alone significantly predicted maternal childrearing attitudes when
mothers’ educational level was controlled. However, a near significant multivariate
interaction of delinquency and minority group status was observed. This trend was most
prominent with respect to maternal childrearing attitudes.

For majority boys, nondelinquent boys had mothers whose childrearing attitudes
were more empathic and who endorsed appropriate family roles. Among minority boys,
nondelinquents had mothers who were less empathic and endorsed greater role reversal
and parentification of the child. The relationship between minority group status and
childrearing attitudes was observed for nondelinquent boys only, while the influence of
mothers’ education level on childrearing attitudes was found primarily for delinquent boys.
No significant relationship was observed between minority group status and childrearing
attitudes for delinquent boys. The findings for majority boys are not surprising in that
substantial research supports a connection between delinquency and cold, harsh maternal
childrearing attitudes. The results pertaining to minority boys were surprising, however.

The results of this study suggest that mothers of nondelinquent minority boys tend to set

71

firmer limits on their sons’ behavior and have higher behavioral expectations of them than
the mothers of delinquent minority boys.

One possible explanation for this unexpected finding is the lack of racial balance
between the delinquent and nondelinquent groups, creating a spurious statistical finding.
An alternative explanation is suggested by research examining influences on delinquency
among minority youth. Several studies have found that existing models of delinquency
inadequately explain delinquent behavior for minority youngsters (Cemkovich &
Giordano, 1987; Gray-Ray & Ray, 1990), and report conflicting findings regarding the
impact of perceived parenting on delinquency. In a sample of Black males, Gray-Ray and
Ray (1990) found parental rejection to be the only significant predictor of serious
delinquency, with parental control and supervision only weakly related to serious
delinquency. Cemkovich and Giordano (1987), however, found that increased intimate
communication between parent and child, thought to be negatively related to delinquency,
was, in fact, positively related to delinquency for minority boys. These authors suggested
that this finding may reflect parent-initiated conversation in which the youth feels
compelled to participate. If this instead reflects more flexible parent-child boundaries, this
finding would be consistent with the current study and explain why increased role reversal
was negatively related to delinquency for minority youth. It would not explain the
apparent relationship between higher levels of empathy and delinquency among minority
youth. Faw and Goldsmith (1980) found that minority adolescents whose parents had
higher expectations of them tended to have fewer behavior problems while majority

adolescents whose parents had high expectations had more behavior problems.

72

An additional explanation may be that the expectation that children will be
responsive and attentive to their parents’ needs serves to keep these boys in closer
proximity to their mothers, and, therefore, more closely attached to them. Bowlby (195 8)
has suggested that the firnction of attachment relationships are to keep the child in close
proximity with a stronger attachment figure who can provide protection. The differences
in childrearing attitudes reported in this study may reflect diverse means by which majority
and minority mothers facilitate an attachment relationship and protect their sons from
engaging in delinquent behavior.

Several hypotheses examined expected difi’erences in selected Rorschach variables
between delinquent and nondelinquent boys. The specific variables examined -- texture
responses, human content responses and human movement responses -- are conventionally
interpreted as representing the individual’s object representational world and providing an
indication of how the individual views interpersonal relationships. It was hypothesized
that delinquent boys would have a greater frequency of “T-less” protocols and fewer
human content and human movement responses, as previous studies have shown that
delinquent boys internalized fewer adequately developed object representations than
nondelinquent boys (Gacono & Meloy, 1994; Weber, 1990; Weber, Gacono, &

Meloy, 1992). No significant differences were found between delinquent and
nondelinquent boys on the three Rorschach variables.

Ward (1966) has suggested that the number of human movement responses may
actually be higher in neglected children due to limitations of spontaneous motor

experience, resulting in a greater reliance upon projection rather than motoric discharge.

73

It seems unlikely, however, that this is the case with delinquent adolescents who engage in
constant motoric discharge. A complicating factor in understanding the Rorschach results
is the inclusion in this study of protocols with fewer than fifteen responses. Such “barren
Rorschachs” are often discarded in research and clinical settings as invalid (Exner, 1986).
Typical of the responses on these “barren” protocols was one boy who responded “Wet
paint. Cause it just looks like it!” to each card. Although technically a valid, scorable
response, Brickman and Lerner (1992) note that structural scoring does not accurately
reflect the personality dynamics of such adolescents. They recommended including a non-
statistical “conceptual” approach to analyzing the Rorschachs of such highly resistant
adolescents.

A unique aspect of this finding is that although the responses of delinquent boys
were similar to the reported results from other studies of conduct disordered and
delinquent adolescents, the nondelinquent sample difi‘ered considerably from other
normative samples of non-patient adolescents reported by Exner (1990). A comparison is
provided in Table 8; however, no statistical tests were conducted due to differences in
sample size and sampling techniques. It is possible that the differences between the results
for adolescents in this investigation and the results reported by Exner (1986) are due to
weaknesses in the sampling methods used by Exner in establishing norms for his
Comprehensive System (Wood, Nezworski, & Stejskal, 1996).

A more plausible explanation is that these results reflect the alienation, isolation,
and detachment of urban adolescents in general. If so, then it becomes even more

important to understand the protective factors which constrain adolescent involvement in

Table 8

74

Comparison of Means and Frequencies AmongGroups

of Adolescents on Select Rorschach Variables

 

 

 

 

 

 

 

 

 

 

 

Pure HumangContent All Human Content

M S_D Freq (°/o) M M2 Freq (°/o)
Delinquent (n=30) 1.77 2.08 19 (63) 4.93 3.62 27 (90)
Nondelinquent (n=30) 1.90 1.50 22 (73) 4.60 2.18 30 (100)
Conduct Disordered (n=100) 2.32 1.95 85 (85) 5.70 3.51 98 (98)
14 year olds (n=105) 3.00 1.71 105 (100) 5.46 2.44 105 (100)
15 year olds (n=110) 3.42 1.96 109 (99) 5.57 2.28 110 (100)
16 year olds (n=140) 3.39 1.94 139 (99) 5.51 2.12 140 (100)
Adult males (n=350) 3.62 1.89 346 (99) 5.72 1.61 350 (100)

Human Movement le_xtu_re

M S_D Freq (°/o) M S_D Freq (°/°)
Delinquent (n=30) 2.50 2.45 23 (77) 0.40 1.48 5 (17)
Nondelinquent (n=30) 2.67 1.71 28 (93) 0.33 0.80 6 (20)
Conduct Disordered (n=100) 3.30 2.43 87 (87) 0.17 0.45 14 ( 14)
14 year olds (n=105) 4.06 2.24 105 (100) 0.99 0.52 85 (81)
15 year olds (n=110) 4.35 2.17 110 (100) 1.06 0.51 101 (92)
16 year olds (n=140) 4.31 2.13 140 (100) 1.02 0.48 128 (91)
Adult males (n=350) 4.54 2.00 350 (100) 0.97 0.50 307 (88)

 

Note. Freq = number of individuals providing at least one response. % = percent of sample

providing at least one response. Conduct disordered adolescent norms from Gacono and Meloy,
1994, Non-patient norms fiom Exner, 1990.

 

75

serious delinquency. These factors may be represented by differences between delinquents
and nondelinquents on other Rorschach indicants such as ego strength and impulse control
(Afr, FC:C+CF) and ability to feel anxiety (Sum Y) (Curtiss, F eczsko, & Marohn, 1979;
Gacono & Meloy, 1994) or by the difl‘erences in verbal ability, early trauma and maternal
childrearing attitudes reported in this study.

Hierarchical regression analyses were employed to examine whether internalized
object representations and perceived quality of the mother-child attachment relationship
could be predicted from early childhood events and maternal childrearing attitudes. The
CLES and AAPI were again employed to measure life events and maternal parenting
attitudes, respectively. Intemalized object relations were measured by combining the
number of human content responses and the number of human movement responses to the
Rorschach Inkblot Test and dividing by the total number of responses. Perceived mother-
child relationship was measured by the Inventory of Parent-Peer Attachment (Armsden &
Greenberg, 1987). Object relations theory proposes that children’s early experiences of
parenting and traumatic life events strongly influence subsequent interpersonal interactions
by contributing to their internalized models of themselves and their relationships with
others. It was hypothesized that boys who experienced harsh parenting and greater
childhood trauma would have poorer internalized object representations. It was also
hypothesized that harsh parenting and greater childhood trauma would influence boys’
perceptions of their relationships with their mothers. These hypotheses were not
supported. Hierarchical regression was also used to test whether maternal attitudes

toward parenting would moderate the effect of early childhood trauma on boys’ object

76

representations and perceived mother-child relationship. No significant relationship was
found between these variables.

Delinquent and nondelinquent boys did not differ in their perceptions of their
relationships with their mothers. In fact, contrary to expectation, delinquent boys scored
in the more positive direction on the EPA than nondelinquent boys. Willock (1986)
suggested that aggressive, antisocial youngsters Struggle to ward off feelings of being
unloved and “utilize massive defensive operations in order to maintain a good image of
mother (p.65)”. Defensive denial in particular is often expressed through
undersocialization and undercontrol of impulse and affect (Shedler, Mayman, & Manis,
1993). Additionally, self report scales such as the EPA have been criticized as
“transparent” and subject to ambiguous interpretation at the healthy or normal end of the
spectrum (Shedler, Mayman, & Manis, 1993). It is also possible that these results reflect
the fact that delinquents who participated in the study had mothers who were more
committed and concerned with their sons, and are therefore less representative of the

mothers of serious delinquents.

Implications for Theory and Practice
This study provides some support for the relevance of attachment theory to
understanding delinquent behavior. Attachment theory postulates that children whose
primary caregivers are not adequately responsive to their needs for safety and security will
respond with increased anger and aggression. The present study confirms this central

precept. Delinquency was associated with a prior history of childhood trauma and, for

77

majority boys, with suboptimal childrearing attitudes. This study was unable to confirm
attachment theory’s position that trauma and parenting influence the development of
internal working models or object representations, or that object representations differed
significantly between delinquent and nondelinquent boys.

Delinquent boys and their mothers responded as expected in this study, with the
exception of a noted group by minority status interaction on the AAPI. The Rorschach
responses of delinquent boys were notably devoid of texture responses and had fewer
human movement and human content responses compared to established adolescent
norms. That nondelinquent adolescents showed a similar lack in these types of responses
is cause for concern about the state of adolescent emotional health and their capacity for
empathy and close interpersonal relationships. Adolescence can be considered a time of
emotional turmoil and upheaval, but these findings point to a deeper, more
characterological rejection of attachment relationships. If this is an accurate reflection of
urban adolescents’ interpersonal firnctioning, it becomes crucial to understand what
protective factors are working to keep these dangerously distant, alienated boys fi'om
engaging in delinquency. Is it merely a question of opportunity? Does attachment to
social institutions such as school provide some protection by acting as a constraint upon
engaging in delinquency, as social control theory suggests (Hirshi, 1969)? The significant
differences in PPVT-R scores and mother’s education level suggests that a family which
places a strong value upon formal education is less likely to have a son involved in
delinquent behavior. This family value may facilitate an attachment between the child and

the school (Hirshi, 1969), increasing the likelihood that the child will experience academic

78

success, and deterring the child from engaging in delinquent behavior. The issue, then,
may be one of (a) creating successful academic experiences for children at risk of firture
delinquency to increase their attachment to schools, and (b) increasing their mothers’
education levels to establish stronger family affiliations with schools.

A clear implication of this study is the need for early intervention services for
young children who have experienced traumatic life events. These children are at risk for
greater physical and emotional health problems than their peers as well as subsequent
involvement in delinquency. It is, therefore, highly appropriate that the school district in
which this study was conducted has embarked upon an ambitious program to provide
counseling services to elementary school students and their parents. Although this
program is not specifically designed as a “delinquency prevention” program, any
significant attempt to help ameliorate childhood trauma and improve family functioning is
bound to ultimately reduce delinquency. Attempts to improve family firnctioning must be
culturally sensitive, understanding ethnic difl‘erences in parenting practices and childrearing

attitudes.

Limitations of the Present Study
This study contains some limitations which may have affected the reSearch
findings. First, it is important to note that only delinquent boys whose mothers came to
visit were eligible to participate in this study. Less than 50% of the youth in detention
during this time received visitors during their incarceration. Furthermore, many of the

youth who received visits had only a father, grandparent or foster parent and were

79

therefore ineligible for this study. Some of the most serious delinquents were either
waived into the adult system or were already remanded into state custody as a result of
severe parental abuse or neglect, and thus were automatically excluded form this study.
The delinquents who participated in this study had mothers who were interested in
participating, many of whom expressed the hope that their participation would contribute
to professional knowledge which might prevent some other child fiom becoming
delinquent. These mothers may have been more involved with their sons, and thus, less
representative of mothers of serious delinquents.

Some long-standing criticisms of delinquency research may be applied to this
current study. Schmidl (1947) noted that “delinquency” is a legal/sociological term, not a
psychological one. Delinquency is a behavioral syndrome which can have various causes.
Delinquents are not a homogeneous group, and it is unlikely, in Schmidl’s estimation, that
any one group could be a representative sample of delinquents. His remarks are echoed
by subsequent investigators such as Quay (1987) and Snyder and Patterson (1987) who
define multiple subgroups of delinquents, each with their own behavioral trajectories and
contributing influences. Agnew (1991) found that influences on delinquency change over
the course of adolescence, with partial attachment of paramount importance in early
adolescence, commitment to school in mid-adolescence, and internal beliefs in late
adolescence. Steinberg (1987) has proposed a developmental model of delinquency which
also delineates differing motivational and other contributing factors.

This study used a quasi-experimental design. By their nature, quasi-experimental

designs contain significant threats to internal and external validity (Cook & Campbell,

80

1979). Obviously, boys could not be randomly assigned to “delinquent” or
“nondelinquent” groups. The inclusion of a nondelinquent comparison group does,
however, diminish some of the threat to this study’s internal validity and is an
improvement on many earlier studies of delinquency which do not include a comparison
group. The most significant threats to external validity involve the samples of delinquent
and nondelinquent boys. Sample sizes in each group were modest. Questions regarding
representativeness of any sample of delinquent youngsters have already been raised. It is
also unclear how representative the comparison group of nondelinquent boys was.
DiLalla, Mitchell, Arthur, and Pagliocca (1988) reported that 80% of American
adolescents admit committing at least one delinquent act. One must then wonder whether
these nondelinquents were truly not involved in criminal behavior or, simply, not involved
in the juvenile justice system. These issues limit the generalizability of this study.
Additionally, this study was restricted to one relatively small geographic location. Policies
regarding who is incarcerated or prosecuted are often influenced by community pressures
and limited intervention resources. Many participants in early studies on delinquency
would barely come to a court’s attention in today’s world of teenage violent crime. It
remains unclear how generalizable these results will remain over time or settings.

This study contains several threats to internal validity. The substantial
intercorrelations of ethnicity, mother’s education level and federal lunch subsidy present
potential risks of multicollinearity and nuisance variables in the regression and

MANCOVA analyses, respectively. The relative effects of maternal education and

81

ethnicity were accounted for statistically; however, it would have been preferable to have
more closely matched groups on these variables.

With regard to instrumentation, it is likely that the results of this study
underestimate the frequency and severity of trauma experienced by many children today.
The CLES assumes that a given event occurs only once during a given time-span. For
example, although the measure includes an item regarding parental divorce or separation,
it assumes that a child will only experience this event once. No provision is made for
children whose parents separate from multiple partners, despite the implications this may
have for the young child’s ability to form stable, trusting relationships. The items
contained in the CLES may not represent the scope and severity of stressful events
experienced by today’s urban youth. For example, it does not contain items regarding
exposure to parental alcoholism, substance abuse, neighborhood violence, or
homelessness. It also does not include being a victim of crime or having a parent who was
a victim of a crime. While it does contain an item regarding parental separation or
divorce, it does not include items regarding witnessing domestic violence, being the target
of a custody dispute or enduring frequent moves. Although research has not
demonstrated a definitive link between these specific types of life stress and subsequent
delinquency, it seems reasonable to suggest that exposure to these events may cause at
least as much trauma and stress as items such as, “Mother began work.” Such events may
be far more common to the life experiences of today’s youth than one would like to

acknowledge.

82

The clinical significance of ethnic differences observed on the AAPI remains
unclear. This may be a statistical artifact of unbalanced groups, or it may signify more
important differences in family dynamics and parental influences on delinquency between
majority and minority boys. It is also unclear what impact, if any, consolidating all
minority boys into a single “Minority” classification may have had, particularly in
understanding cultural differences in maternal childrearing attitudes. A firrther limitation
of the CLES and AAPI are their retrospective nature. This study’s reliance upon maternal
retrospective reports of childrearing attitudes and experience of stressfirl life events
presented several problems. Mothers may have under-reported the occurrence of events
or inaccurately recalled how old their sons were when these events occurred. During the
data collection process, many mothers remarked, “I don't remember exactly how old he
was when this happened.” The limitations of the CLES itself have already been discussed.
Second, mothers’ attitudes toward childrearing may have changed or may not reflect
actual parenting behavior. One mother of a delinquent boy endorsed “Strongly Disagree”
to the AAPI item “Children should always be spanked when they misbehave” while
stating, “I don’t believe that anymore. I beat both my boys when they were young. Now
the big one’s in jail and the little one’s here (juvenile detention), so I guess that don’t
work.” An alternative approach such as a longitudinal study or in-depth interview might

have resulted in greater differences in trauma and maternal childrearing attitudes.

83

Recommendations for Future Research

 

The results presented in this study are best understood as a preliminary
investigation. Further replication with larger samples is needed to further delineate the
complex, multifaceted ways in which trauma and parenting both contribute to delinquency.
A larger sample size would have numerous benefits. The sample of delinquent youth
could be further divided into relevant subgroups. As delinquents are not a homogenous
group, it would be usefirl to distinguish such categories as aggressive and non-aggressive,
gang-involved and solitary, recidivists and first offenders, as well as having more culturally
diverse delinquent and non-delinquent groups. Such distinctions might be made through
the inclusion of formal interviews such as the Hare Psychopathy Checklist to assess
antisocial traits or the Diagnostic Interview Schedule for Children to assess for psychiatric
disorders. Furthermore, the presence of a larger sample would allow for the inclusion of
more Rorschach variables and the exclusion of protocols with questionable validity. An
increased sample Size would also increase the power of statistical analyses, especially
regression analyses, and maximize the likelihood of detecting significant statistical
relationships.

Future research must include comparison groups that are well matched with
respect to age, ethnicity, socioeconomic status and parental education level. These factors
were found to be highly intercorrelated with delinquency as well as parental childrearing
attitudes and early childhood trauma. Such relationships often obscure rather than
illuminate and present methodological complications which can only be partially addressed

by statistical means. Furthermore, the results of this study hint at important differences

84

between majority and minority boys. Certainly the inclusion of more racially balanced
groups would permit more ethnically-sensitive statements about differential paths to
delinquency.

The results of this study have limited generalizability. It will be important to
replicate these findings with adolescents in other geographical locations. It will also be
important to include females to discern differences in male and female pathways to
delinquency. The current investigation employs a comparison group of nondelinquent
boys which is an improvement over many previous studies; however, future studies might
include a more narrowly defined population of adolescents who show strong prosocial
attitudes such as youth involvement in community activities or volunteer work rather than
an absence of criminal history. While the focus of this study was on mother-child
relationships, it would be usefirl to examine the salience of the father-child relationship in
firture studies as well.

While significant differences were found in frequency and severity of early
childhood trauma, the present finding was likely an underestimate of frequency, severity,
and type of trauma experienced by contemporary adolescents. Future research might
employ a more current measure of traumatic events which accounts for the limitations of
the CLES. It would also be of interest to examine parent-chfld concordance on such a
measure. Such an investigation might provide some indication as to whether the
experience of trauma or the individual’s perception of the trauma has a more profound
effect on subsequent functioning. With regard to the use of projective testing as an tool

for examining internal object representations, it remains unclear why the nondelinquent

85

sample performed so similarly to the delinquent sample and to clinical samples reported in
earlier studies. Areas for fixture study might include the scoring of additional Rorschach
variables or the inclusion of more comprehensive object representation scales such as
those of Blatt (Blatt & Lerner, 1983; Blatt, Brenneis, Schiemek, & Glick, 1976) or
Kwawer (1979, 1980), or the use of other projective instruments such as the Thematic
Apperception Test.

The final recommendation is for more longitudinal research to determine the
direction of the relationship between variables such as parenting attitudes and delinquency.
Liska and Reed (1985) have shown the importance of cross-lagged studies in
demonstrating the causal relationship between parenting and delinquency. A longitudinal
study might also include observation of actual parent-child interaction at several points
throughout childhood and adolescence.

Juvenile delinquency is a complex, multifaceted problem that has long perplexed
psychologists, sociologists, criminologists and educators. While many studies have sought
to examine contributing factors to delinquency, what has often been lacking in prior
research has been a comprehensive, unifying theoretical framework. By utilizing tenets of
attachment theory, this study examined the relative contributions of childhood trauma and
maternal childrearing attitudes. The exact means by which these factors influence
subsequent delinquent behavior were not successfully clarified by the current study. It is
hoped that future research will be able to demonstrate the influence of trauma and

parenting on the formation of object representations and subsequent delinquent behavior.

APPENDICES

86

Appendix A
Parental Consent Form

I am interested in learning about factors that afi’ect teenage behavior in order to
improve education and other services for teens. You have been identified as the mother of a
teenage boy. I invite you and your son to participate in a study I will be conducting. If you
agree to participate, you will be asked to complete several brief questionnaires. Your son will
be asked to complete several measures of teenage behavior.

The parent questionnaires will take a total of about 30 minutes to finish. The teen
measures will take about 90 minutes to complete and will be given during the school day by
advanced doctoral students in psychology. Your participation and your teenager’s participation
are completely voluntary. You and your teenager may choose to not answer any question or
stop participating at any time. Consent or refusal to participate will not affect services
provided to your or your son in any way.

Answers to all measures will be kept confidential. Summaries of the information
gathered and statistics used will be based on group data, but will not identify individuals. A
summary of the research results may be obtained upon the study's completion by contacting
me.

I hope that you will choose to participate and allow your son to participate. Please feel
free to contact me if you have any additional questions.

Yours truly,
Sandi Isaacson, MEd.

Please Sign both statements.

I understand and agree to participate in this study.

 

 

Signature Date

I agree to allow my child, , to participate in this study.

 

 

 

Signature Date

87

Appendix A (continued)

Youth Assent Form

Verbal instructions to be read to teenager PRIOR to testing:

We would like to learn more about teenage behavior and would like you to be part of
our study. The study will take about 1 1/2 hours. I will ask you some questions and show you
some drawings and designs.

You do not have to participate if you don‘t want to. If you decide to participate, you
can stop at any time, or choose not to answer any question you feel uncomfortable with. All of
your answers will be kept confidential. They will not be shared with your parents, teachers or
anyone else except experimenters conducting this study. Your name will be removed fi'om all
answer sheets.

Whether or not you participate is up to you. Your decision will not affect your [grades
at school] / [treatment or length of stay at the Youth Center] (read appropriate phrase). If you
or your parents have any questions, you can ask me or your parents can call Ms. Isaacson, who
is running this study, at XXX-XXXX.

Do you have any questions?

Would you like to be in this study?

88

Appendix B
GENERAL INFORMATION QUESTIONNAIRE

In order to help us get a better "picture" of your son, please answer the following brief
questions. All responses will be kept confidential.

Child's age

 

Current grade

 

Race/ethnicity

 

Last grade of school yo_u completed

 

Do you receive ADC or Medicaid?

 

Thank you for your participation in this study.

89

Appendix C

Children's Life Events Questionnaire
(Coddington, 1972)

Instructions

Listed below are some events they your child may have experienced during his first four
years. Please check the events that your child experienced. Leave it blank if your child
did not experience the event.

 

 

 

 

 

 

 

_ 2.
* 3.
* 4.
* 5.
___6.
* 7.
* 8.
__ 9.
* 10.
____11.
_ 12.
_13.
__14.
* 15.
* 16.

 

Child began nursery, preschool, or daycare
Child changed to a different school or daycare
Birth or adoption of a brother or sister
Brother or sister left home

Hospitalization of brother or sister

Death of brother or sister

Change of father’s job requiring increased absence from home
Loss of job by parent (mother/father)

Marital separation of parents

Divorce of parents

Hospitalization of parent (serious illness)
Death of a parent

Death of a grandparent

Marriage of parent to stepparent

Jail sentence of parent for 30 days or less

Jail sentence of parent for 1 year or more

90

Appendix C (continued)

* 17. Addition of third adult to family (e. g., grandparent)

 

18. Change in parents’ financial status

* 19. Mother began work

 

20. Decrease in number of arguments between parents

21. Increase in number of arguments between parents

22. Discovery of being an adopted child

* 23. Acquiring a visible deformity

 

* 24. Having a visible congenital deformity

 

* 25. Child hospitalized

 

* 26. Change in acceptance by peers

 

27. Outstanding personal achievement

* 28. Death of child's close friend

 

Note. * = items on which delinquents differed significantly from nondelinquents (p < .05).

91

Appendix D
EPA

(Armsden & Greenberg, 1987)

Teenagers have many different feelings about their mothers. Please answer each item as
honestly as you can. All responses will be kept confidential; they will not be shared with your
parents, teachers or classmates.

” ‘6

Indicate whether the following items are “always or almost always true , often true”,

’9 “

“sometimes true , seldom true”, or “almost never or never true” for you.

Never Seldom Sometimes Often Always

True True True True True
My mother respects my feelings. I 2 3 4 5
I feel my mother is successful 1 2 3 4 5
as a parent.
I wish I had a different mother. 1 2 3 4 5
My mother accepts me as I am. 1 2 3 4 5
I have to rely on myself when I 1 2 3 4 5
have a problem to solve.
I like to get my mother’s point of 1 2 3 4 5
view on things I’m concerned about.
I feel it’s no use letting my 1 2 3 4 5
feelings show.
My mother senses when I’m upset I 2 3 4 5

about something.

92

Appendix D (continued)
Never Seldom Sometimes Often Always
True True True True True

Talking over my problems with my 1 2 3 4 5
mother makes me feel ashamed or foolish.

My mother expects too much from me. 1 2 3 4 5
I get upset easily at home. 1 2 3 4 5
I get upset a lot more than my 1 2 3 4 5

mother knows about.

When we discuss things, my mother 1 2 3 4 5
considers my point of view.

My mother trusts my judgment. 1 2 3 4 5
My mother has her own problems, 1 2 3 4 5
so I don’t bother her with mine.

My mother helps me to understand 1 2 3 4 5
myself better.

I tell my mother about my problems 1 2 3 4 5
and troubles.

I feel angry with my mother. 1 2 3 4 5
I don’t get much attention at home. 1 2 3 4 5
My mother encourages me to talk 1 2 3 4 5
about my difliculties.

My mother understands me. 1 2 3 4 5
I don" know whom I can depend 1 2 3 4 5

on these days.

When I am angry about something, 1 2 3 4 5
my mother tries to be understanding.

I trust my mother.

My mother doesn’t understand
what I’m going through these days.

I can count on my mother when I
need to get something off my chest.

I feel that no one understands me.

If my mother knows something is
bothering me, she asks me about it.

93

Appendix D (continued)
Never Seldom Sometimes Often Always
True True True True True
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
l 2 3 4 5
1 2 3 4 5

REFERENCES

REFERENCES

Acklin, M. W., Bibb, J. L., Boyer, P., & Jain, V. (1991). Early memories as expressions of
relationship paradigms: A preliminary investigation. JoumflfPersonalipy Assessment,
fl, 177-192.

Agnew, R (1991). A longitudinal test of social control theory and delinquency. Journal of
Research in Crimefi and Delinquency a, 126-156.

 

Ainsworth, M. D. (1964). Pattems of attachment behavior shown by the infant in interaction
with his mother. Merrill-Palmer Quarterly, m, 51-58.

 

Ainsworth, M. D. (1969). Object relations, dependency and attachment: A theoretical review
of the infant-mother relationship. gild Developm, fl, 969-1025.

Ainsworth, M. D. (1979). Infant-mother attachment. American Psychologist, 3_4, 932-937.
Ainsworth, M. D. (1985). Patterns of attachment. The Clinical chhologiat, 38, 27-29.
Ainsworth, M. D. (1989). Attachments beyond infancy. American Psychologiat, fl, 709-716.

Allen, J. P., Aber, J. L., & Leadbeater, B. J. (1990). Adolescent problem behaviors: The
influence of attachment and autonomy. Psychiatric Clinjcs of North America, Q,
455-467.

Annsden, G. C., & Greenberg, M. T. (1987). The Inventory of Parent and Peer Attachment:
Individual difi‘erences and their relationship to psychological well-being in adolescence.
Jomal of Youth and Adolescence, _1_6_, 427-454.

Athey, G. 1., Fleicher, J., & Coyne, L. (1980). Roschach object representation as influenced by
thought and afi’ect organization. In J. S. Kwawer, H. D. Lerner, P. M. Lerner, & A
Sugarman (Eds). Borderline phenomena_ and the Rorschach test (pp. 275-278). New
York: International Universities Press.

 

94

95

Baron, R M., & Kenny, D. A (1986). The moderator-mediator variable distinction in social
psychological research: Conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, fl, 1173-1182.

Bavolek, S. J. (1984). Iiandbmlg for the Adult-Adolescent Parenting Inventory (AAPI).
Shaumberg, FL: Family Development Associates, Inc.

Bene, E. (1975). An effect on Rorschach M responses of a boy's relationship with his mother.
Journal ofPersonality Assessmant, 3, 114-115.

Birigin, Z. (1994). Attachment theory and research: Application to clinical practice.
American Journal of Orthopsychm, Q, 404-420.

Blatt, S. J ., & Lerner, H. D. (1983). The psychological assessment of object representation.
J oumal of Personality Assessmen_t, fl, 7-28.

Blatt, S. J ., & Berman, W. H. (1984). A methodology for the use of the Rorschach in clinical
research. Journal of Personality Assessment, 4_8, 226-239.

Blatt, S. J ., Brenneis, C. B., Schimek, J. G., & Glick, M. (1976). Normal development and
psychopathological impairment of the concept of the object on the Rorschach. Journal
of Abnormal Psychology, 85(4), 364-3 73.

Blatt, S. J., Tuber, S. B., & Auerbach, J. S. (1990). Representation of interpersonal
interactions on the Rorschach and level of psychopathology. Jouma of Personalm
Assessment, 3, 711-728.

Bowlby, J. (1947). 44 Juvenile thieves: Their cflacteriatics and home life. London: Balliere,
Tindall & Cox.

Bowlby, J. (195 8). The nature of the child's tie to his mother. International Journal of
Pachoaaagsis, 3_9, 350-373.

Bowlby, J. (1977a). The making and breaking of afl‘ectational bonds: I Aetiology and
psychopathology in the light of attachment theory. BritifiJoufla] of Psychm m,
201-210.

Bowlby, J. (1977b). The making and breaking of afi‘ectational bonds: II Some principles of
psychotherapy. Briti_sh Joumal of Psychiatry, 11C), 421-431.

Bowlby, J. (1988). A secure basi Parent-child attachment and healm human development.
New York: Basic Books.

96

Brand, H. J ., Crous, B. H., & Hanekorn, J. D. (1990). Perceived parental inconsistency as a
factor in the emotional development of behavior-disordered children. Pychological
Ramrts, 66, 620-622.

Bray, J. H., & Maxwell, S. E. (1982). Analyzing and interpreting significant MANOVAs.
Review of Educational Research, :2, 340-367.

Brennan, P. A, & Mednick, S. A. (1990, August). Childhood psychopathology as a predictor
of violen_t criminal behavior. Paper presented at the 98th Annual Convention of the
American Psychological Association, Boston, MA.

 

Brickman, AS, & Lemer, H. D. (1992). Barren Rorschachs: A conceptual approach.
Journal of Personality Assesament. 52, 176-184.

Burbach, D. J., Kashani, J. H., & Rosenberg, T. K. (May 1989). Parental bonding and
depressive disorders in adolescents. Journal of Ch_ild Psychology & Psychiatry &
Allied Disciplineg, A), 417-429.

Call, J. D. (1984). Child abuse in infancy: Sources of hostility within the parent-infant dyad
and disorders of attachment in infancy, Child Abuse aad Neglect, 8, 185-202.

 

Carison, V., Cicchetti, D., Barnett, D., & Braunwald, K. (1989). Disorganized/disoriented
attachment relationships in maltreated infants. Developmeng Psychology, 2, 525-
531.

Cemkovich, S., & Giordano, P. (1987). Family relationships and delinquency. Criminology,
2; 295-321

Cicchetti, D., & Barnett, D. (1992). Attachment organization in maltreated preschoolers.
DevelopmenL and Psychopatholggy, ;, 397-411.

 

Coddington, R. D. (1972a). The significance of life events as etiologic factors in the diseases
of children, I: A survey of professional workers. Journal of Psychosomatic Research,
as, 167-181.

Coddington, R D. (1972b). The significance of life events as etiologic factors in the diseases
of children, H: A study of a normal population. Journal of Psychosomatic Resear_ch,
1a, 205-213.

Coddington, R D. (1984). Measuring the stressfirlness of a child's environment. In J. H.
Humphrey (Ed), Stres_s in Chiflhood (pp. 97-126). New York: AMS Press, Inc.

Cohn, D. A(1990). Child-mother attachment of six-year—olds and social competence at
school. @d Development. 91, 152-162.

97

Cook, T. D., & Campbell, D. T. (1979). '- ' entation. New York: Rand McNally.

Crittenden, P. M. (1990). Internal representational models of attachment relationships. Infant
Mental Health Journal, 1_1_, 259-276.

Crowell, J. A, O’Connor, E., Wollmers, G., Sprafldn, J ., & Rao, U. (1991). Mothers’
conceptualizations of parent-child relationships: Relation to mother-child interaction
and child behavior problems. Development__ and Psychopathology, 8, 431-444.

Curtiss, G., Feczko, M. D., & Marohn, R C. (1979). Rorschach differences in normal and
delinquent white male adolescents: A discriminant fimction analysis. Journal of Youth
and Adolescence, 8(4), 379-392.

Davidow, S., & Bruhn, A R. (1990). Earliest memories and the dynamics of delinquency: A
replication study. Joumal of Personality Assessment, 5_4, 601-616.

DeHorn, A, & Klinge, V. (1978). Correlations and factor analysis of the WISC-R and the
Peabody Picture Vocabulary Test for an adolescent psychiatric sample. J oumal of
Consulting and Clinical Psychology, 56, 1160-1161.

Deutsch, L., & Erickson, M. T. (1989). Early life events as discriminators of socialized and
undersocialized delinquents. J oumal of Abnormal Child chhology, 11, 541-551.

DeVos, G. A (1979). A Rorschach comparison of delinquent and non-delinquent Japanese
family members. Journal of Psychological Anthropology, _2_, 425-441.

DiLalla, L. R, Mitchell, C. M., Arthur, M. W., & Pagliocca, P. M. (1988). Aggression and
delinquency: Family and environmental factors. J oumal of Youth apfldolescence,
1_7_, 233-246.

Dodge, K. A, Pettit, G. 8., & Bates, J. E. (1994). Socialization mediators of the relation
between socioeconomic status and conduct problems. _C_laild Development, _6_5_, 649-
655.

Dunn, L. M., & Dunn, L. M. (1981). Paabody Picture Vocabulary Teat - Revigg. Circle
Pines, MN: American Guidance Service.

Endacott, J. L. (1941). The results of 100 male juvenile delinquents on the Rorschach Inkblot
Test. Journal of Crim'aal Psychopatholagy, 8, 41-50.

Exner, J. E., Jr. (1986). The Rorschach: A comprehensive system. Volume 1: Baaig
Foundations, (2nd ed.). New York: John Wiley & Sons, Inc.

98
Exner, J. E., Jr. (1990). A Rorschach workboak for the comprehenaive system (3rd ed.)
Asheville, NC: Rorscharch Workshops.

Exner, J. E, Jr. (1991) ). The Rorschach: A comprehenaive system. Volume 2:
Intgpretation, (2nd ed). New York: John erey & Sons, Inc.

 

Exner, J. E., Jr., Bryant, E., & Mrller, A (1975). Rorschach resmnses of some juven_iI_e
ofl‘enders. Unpublished manuscript. Asheville, NC: Rorschach Workshops.

Exner, J. E., Jr., Thomas, E. A, & Mason, B. J. (1985). Children”s Rorschachs: Description
and prediction. Journal of Personality Assessment, fl, 13-20.

Exner, J. E., Jr., Vrglione, D. J ., & Gillespie, R (1984). Relationships between Rorschach
variables as relevant to the interpretation of structural data. Journal of Personm
Assessment, 4_8, 65-70.

Farrington, D. P., (1978). The family backgrounds of aggressive youths. In L. Hersov, & M.
Berger (Eds), Aggtession and anti-social behavior in children and adolescents (pp.73-
93). New York: Pergamon Press.

Faw, T. T., & Goldsmith, D. F. (1980). Interpersonal perceptions within families containing
behavior problem adolescents. Journal of Youth and Adolescence, 2, 553-556.

Federal Bureau of Investigation (1995). Uraiform Crime Report. Washington, DC:
Government Printing Office.

Feehan, M., McGee, R, Stanton, W. R, & Silva, P. A (1991). Strict and inconsistent
discipline in childhood: Consequences for adolescent mental health. British Journal of
Clinical Pychology, 3_(), 325-331.

Fishler, P. H., Sperling, M. B., & Carr, A. C. (1990). Assessment of adult relatedness: A
review of empirical findings fi'om object relations and attachment theories. Journal of
Personality Assessment, 88, 499-520.

Freud, S. ([1905]/1935). Three essays on the theory of sexuality. Standard Edition, 1, 130-
245.

Fritsch, R C., & Holstrom, R W. (1990). Assessing object representations as a continuous
variable: A modification of the Concept of the Object on the Rorschach scale. Journal
of Personality Assessment, 88, 319-334.

Furman, R A, & Furman, E. (1984). Intermittent decathexis: A type of parental dysfunction.
International Journal of Psychoanalysis, 6_5, 423-433.

99

Gacono, C. (1988). A Rorschach analasis of ogect relations and defensive structure an_d their

relationship to narcissism and pagchopathy in a gtoup of antisocial ofl‘enders.
Unpublished doctoral dissertation, United States International University, San Diego,

CA

Gacono, C. B. (1990). An empirical study of object relations and defensive operations in
antisocial personality disorder. Journal of Personality Assessment, 5_4, 589-600.

Gacono, C., & Meloy, J. R. (1991). A Rorschach investigation of attachment and anxiety in
anti-social personality disorder. Journal of Nervous & Mental Disease, 1_7_9, 546-552.

 

Gacono, C., & Meloy, J. R. (1994). The Rorschach and assessment of aggtessive and
pgchopathic personaaties. Hillsdale, NJ: Lawrence Erlbaum.

Gold, M., & Mann, D. (1972). Delinquency as a defense. American Joumflpf
Orthopggchiag, fl, 463-479.

Goldberg, S. (1988). Risk factors in attachment. Canadian Journal of Psychology, 42,, 173-
188.

Goldberg, S. (1991). Recent development in attachment theory and research. Canadian
Journg of Psychiatry, 3_6, 393-400.

 

Goldfarb, W. (1944). Effects of early institutional care on adolescent personality: Rorschach
data. American Jouma of Orthopsychia_try, 15, 441-447.

Gove, W. R, & Crutchfield, R. D., (1982). The family and juvenile delinquency. Sociological
MIX, 2_3, 301-319-

Gray-Ray, P., & Ray, M. (1990). Juvenile delinquency in the Black community. Youth and
Social, _2_2, 67-84,

Greco, C. M., & Cornell, D. G. (1992). Rorschach object relations of adolescents who
committed homicide. Journal of Personality Assessment, 5_9, 574-583.

Haapasalo, J ., & Tremblay, R E. (1994). Physically aggressive boys fiom ages 6 to 12:
Family background, parenting behavior, and prediction of delinquency. J oumal of
Consultingand C1ini_cal Psychology, Q, 1044-1052.

 

 

Haase R F ., & Ellis, M. V. (1987). Multivariate analysis of variance. JoumalafCounaeLng
chhology, 3, 404-413.

100

Heisel, J. S., Rean, S., Raitz, R, Rappaport, M., & Coddington, R D. (1973). The
significance of life events as contributing factors in the diseases of children. Journal of
Pediatrics, 83, 119-123.

Hirshi, T. (1969). Causes of delinquengy. Berkeley: Free Press.

Holaday, M., Annsworth, M. W., Swank, P. R, & Vrncent, K. R. (1992). Rorschach
responding in traumatized children and adolescents. Journal of Traumatic Stress, 8,
119-129.

Hopkins, J. (1990). The observed infant of attachment theory. British Journal of
Pyohotherapy, 8, 460-470.

Homer, A J. (1984). Object relations aad the developing ago in therapy. Northvale, NJ:
Jason Aronson Inc.

Homer, A J. (1991). chhoanalytic obiaect relations thmpy. Northvale, NJ: Jason Aronson,
Inc.

Howard, J. (1981). The expression and possible origins of depression in male adolescent
delinquents. Australian and New Zealand Journal of Pyohiatry, 1_5, 311-318.

Johnson, R E. (1987). Mother’s vs. father’s role in causing delinquency. Adolescence, Q,
305-3 15.

Kendall, P.C., & Little, V. L. (1977). Correspondence of brief intelligence measures to the
Wechsler scales with delinquents. Journal of Conarltiagand Clinical chholgy, 4_5,
660-666.

 

Kestenbaurn, C. J. (1984). Pathological attachments and their relationship to afl‘ective
disorders in adult life. American Journal of Psycho is, g, 33-49.

King, C. H. (1976). The ego and the integration of violence in homicidal youth. Annual
Progess in Child Psychiafl and Development, 695-709.

Klopfer, B., Ainsworth, M. D., Klopfer, W. G., & Holt, R R (1954). Development in the
Rorschaah techaiaue. Volume 1. Technique and theoty. Yonkers-on Hudson, NY:
World Book.

Kobak, R R, & Sceery, A. (1988). Attachments in late adolescence: Working models, afl‘ect
regulation, and representations of self and others. Child Developmera, 2, 135-146.

101

Kolvin, 1., Miller, F. J., Fleeting, M., & Kolvin, P. A (1988). Social and parenting factors
affecting criminal ofiense rates: Findings fiom the Newcastle Thousand F arnily Study
(1947-1980). British Jom of Psychim, m, 80-90.

 

Kwawer, J. S. (1979). Borderline phenomena, interpersonal relations, and the Rorschach test.
Bulletin of the Menning Clinic, 48, 515-524.

Kwawer, J. S. (1980). Primitive interpersonal modes, borderline phenomena, and Rorschach
content. In J. S. Kwawer, H. D. Lerner, P. M. Lerner, & A Sugarman, (Eds).
Borderline Phenomena _and the Rorschamest (pp. 89-105). New York: International
Universities Press, Inc.

Levine, L. B. & Tuber, S. B. (1993). Measures of mental representations: Clinical and
theoretical considerations. Bulletin of the Menninger Clinic, 2, 69-87.

Lewis, D. O. (1992). From abuse to violence: Psychophysiological consequences of
maltreatment. J oumal of the American Acadmy of Child and AdolesceaPaychiaaay,
3_1_, 383-391.

Lewis, D. O., Lovely, R, Yeager, C., & Femina, D. D. (1989). Toward a theory of the genesis
of violence: A follow-up study of delinquents. Jourpal of the American Academy of
Child and Adolescent Pym, E, 431-436.

Lewis, D. O., Pincus, J. H., Bard, 8., Richardson, E, Prichep, L. S., Feldman, M., & Yeager,
C. (1988). Neuropsychiatric, psychoeducational and family characteristics of 14
juveniles condemned to death in the United States. American Journal of Patchiatry,
14_5, 585-589.

 

Lewis, D. O., Shanock, S. S., Pincus, J. H., & Glaser, G. H. (1980). Violent juvenile
delinquents: Psychiatric, neurological, psychological and abuse factors. Annual
Progtess in Ch_ild Psychiat_ry and Developmeat, 591-603.

Liska, A E., & Reed, M. D. (1985). Ties to conventional institutions and delinquency:
Estimating reciprocal efi‘ects. American Sociolog'cal Review, 80, 547-560.

Loeber, R, & Dishion, T. (1983). Early predictors of male delinquency. chhological
Bulletin, 9_4, 68-99.

Loeber, R, & Stoutharner-Loeber, M. (1986). Family factors as correlates and predictors of
juvenile conduct problems and delinquency. In M. Tonry, & N. Morris (Eds), an:
and Justice: A review of resaarch. Volume 12 (pp. 29-149). Chicago: University of
Chicago Press.

102

Longfellow, C., & Belle, D. (1984). Stressful environments and their impact on children. In J.
H. Humphrey, (Ed). Stress in Childhood. (pp. 63-78). New York: AMS Press, Inc.

Lopez, F. G., & Gover, M. R (1993). Self-report measures of parent-adolescerrt attachment
and separation-individuation: A selective review. Journal of Counseling and
Development, 21, 560-569.

Loss, N., Beck, S. J., & Wallace, A (1995). Distressed and nondistressed third- and sixth-
grade children’s self-reports of life events and impact and concordance with mothers.
Jom of Abnormal Child Psychology, 3, 397-409.

Luntz, B. K. & \Vrdom, C. S. (1994). Anti-social personality disorder in abused and neglected
children grown up. American Journal of Psychm, ifl, 670—674.

Lyons-Ruth, K., Alpern, L., & Repacholi, B. (1993). Disorganized infant classification and
maternal psychosocial problems as predictors of hostile-aggressive behavior in the
preschool classroom. Claild Development, 84;, 572-585.

 

Maguire, K., & Flanagan, T. J. (1991). Sourcebooic ofcrimiaaljustice statistics 1990.
Washington DC: US. Department of Justice, Bureau of Justice Statistics, US.
Government Printing Office.

Mahler, M. S. (1979). Selected papers of Margaret S. Mfler. Volames I & H. New York:
Jason Aronson, Inc.

Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood and adulthood: A
move to the level of representation. In I. Bretherton & E. Waters (Eds), Growing
paints of attachment theory aflreseaLh. Monographs of the Society for Research in
C_hrl'__d Development, 88, 1-2, Serial #209 (pp. 66-104).

 

Main, M., & Solomon, J. (1986). Discovery of an insecure-disorganized attachment pattern.
In T. B. Brazelton & M. W. Yogrnan (Eds), Afl‘ective develognent in infang. (pp.
95-124). Norwood, NJ: Ablex Publishing Corp.

Main, M., & Stadtrnan, J. (1981). Infant response to rejection of physical contact by the
mother: Aggression, avoidance, and conflict. J oumg of the AmericaaAcademyof
Child and Adolescent Psychm, 29, 292-307.

Majumber, A K., & Roy, A. B., (1962). Latent personality content of juvenile delinquents.
Journal of Psychological ReseaLh, 1, 4-8.

Mak, A S. (1990). Testing a psychological control theory of delinquency. Criminal Justice
and Behavior, 12, 215-230.

103

Mak, A S. (1991). Psychosocial control characteristics delinquents and non-delinquents.
Criminal Justice and BehLvior. 18., 287-303.

Maymam, M. (1967). Object representations and object relationships in Rorschach responses.
Journal of Projective Techniques and Personalig Assessment, 81, 17-24.

McCrone, E. R, Egeland, B., Kalkoske, M., & Carlson, E. A (1994). Relations between early
maltreatment and mental representations of relationships assessed with projective
storytelling in middle childhood. Development__ and Psychopatholctgy, 8, 99-120.

Meloy, J. R (1988). The psychopathic mind: Origina' , dmp'cs, and treatment. Northvale,
NJ: Jason Aronson, Inc.

Montalto, F. D. (1952). Maternal behavior and child personality: A Rorschach study. Journal
of Projective Technniques. 1g, 151-178.

Nada Raja, S., McGee, R, & Stanton, W. R (1992). Perceived attachments to parents and
peers and psychological well-being in adolescence. Journal of Youthanda‘idolescence,
2_1, 471-485.

Parkes, C. M. (1982). Attachment and the prevention of mental disorders. In C. M. Parkes &
J. Stevenson-Hinde (Eds), The plzae of attachment in human behavior. (pp. 295-309).
New York: Basic Books.

Paterson, R 1., & Moran, G. (1988). Attachment theory, personality development and
psychotherapy. Clinical PsychOIOJy Review. 8, 611-636.

Pierce, G. (1978). The absent parent and the Rorschach "T" response. In E. J. Hunter and D.

Nice (Eds). Children of milim families (Chapter 4). Washington DC: US.
Government Printing Office.

Quay, H. C. (1987). Patterns of delinquent behavior. In H. C. Quay (Ed), Handbook of
juvenile delinqaency (pp. 118-138). New York: John Wiley, Inc.

Ray, A (1963). Juvenile delinquency by Rorschach inkblots. Paychologia, 8, 190-192.

Raychaudhuri, M., & Maitra, A K. (1965). A Rorschach study of normal, convicted and
incipient delinquent adolescents. Rorschachiana J amnica, 8, 17 5-189.

Renken, B., Egeland, B., Marvinney, D., Mangelsdorf, S., & Sroufe, S. (1989). Early
childhood antecedents of aggression and passive withdrawal in early elementary school.
burial of Personaliy, 5_7, 257-281.

104

Rey, J. M., & Plapp, J. M. (1990). Quality of perceived parenting in oppositional and conduct
disordered adolescents. J oumal of the American Aaademy of Child and Adolescent
Paychiag, 2_9, 382-385.

 

Rice K. G. (1990). Attachement in adolescence: A narrative and meta-analytic review.
Joum of Youth and AQolescence, 1_9, 511-538.

Rorschach, H. (1942). Psychodragnostics. New York: Grune & Stratton. (Original work
published 1921).

Sampson, R J ., & Laub, J. H. (1994). Urban poverty and the family context of delinquency: A
new look at structure and process in a classic study. Ch_ild Development, Q, 523-540.

Sandler, I. N., & Block, M. (1979). Life stresses and maladaption of children. American
Journal of Communin chhology, 1, 425-440.

Schachtel, E. G. (1951). Notes on Rorschach tests of 500 juvenile delinquents and a control
group of 500 non-delinquent adolescents. Journal of Projection Techpolagy, Q, 144-
72.

Schmidl, F. (1947). The Rorschach test in juvenile delinquency research. American Journal of
Orthopglchiag, ll, 151-160.

Shedler, J ., Mayman, M., & Mannis, M. (1993). The illusion of mental health. American
Paychologist, 4_8, 1117-1131.

Sirnons, R L., Robertson, J. F ., & Downs, W. R. (1989). The nature of the association
between parental rejection and delinquent behavior. Journal of Youth and
Adolescence, 1_8, 297-310.

Sinha, B., Singh, A K., & Singh, S. K. (1986). Delinquents: An observation of their
Rorschach protocols. Indian Paychological Review, 89, 4-8.

Snyder, J ., & Patterson, G. (1987). Family interaction and delinquent behavior. In H. C. Quay
(Ed). Handbook of juvenile delinquengj. (pp. 216-243). New York: John erey &
Sons, Inc.

Spear, W. E. (1980). The psychological assessment of structural and thematic object
representations in borderline and schizophrenic patients. In J. S. Kwawer, H. D.
Lerner, P. M. Lerner, & A. Sugarman (Eds), Borderline phenomana_ and the
Rorschach test. (pp, 321-340). New York: International Universities Press, Inc.

Speiker, S. J. (1986). Patterns of very insecure attachment found in samples of high risk
infants and toddlers. Topics in Baaly Childhood Special Education. 8, 37-53.

105

Speltz, M. L., Greenberg, M. T., & Deklyen, M. (1990). Attachment in preschoolers with
disruptive behavior: A comparison of clinic-referred and non-problem children.
Development and Psychopathology, 2, 31-46.

Sroufe, A (1982). Infant-caregiver attachment patterns of adaptation in preschool: The roots

of maladaption and competence. In M. Perlrnutter, (Ed), Minnesota msium in
child pgchology. (pp. 41-83). Minneapolis, MN: University of Minnesota Press.

Sroufe, L. A, & Waters, E. (1977). Attachment as an organizational construct. Child
Development. :8, 1184-1199.

Stark, S. (1966). Role-taking, empathetic imagination and Rorschach human movement
responses: A review of two literatures. Perceptual_ and Motor Sfilla, Q, 243 -256.

Stayton, D. J. & Ainsworth, M. D. S. (1973). Individual difl‘erences in infant responses to
brief, everyday separations as related to other infant and maternal behaviors.
Developmeng Psychology, 2, 226-235.

Steinberg, L. (1987). Familial factors in delinquency: A developmental perspective. Journal of
~Adolescent Researgr, 2, 255-268.

Stricker, G., & Healey, B. D. (1990). Projective assessment of object relations: A review of
the empirical literature. Paychological Assessment: A Joumfi of Conadfim
Clinical Pychology, 2, 219-230.

Tracy, R L., & Ainsworth, M. S. (1981). Maternal afl‘ective behavior and infant-mother
attachment patterns. C_hild Developmera, 82, 1341-1343.

Van Voorhis, P., Vullen, F. T., Mathers, R. A, & Garner, C. C. (1988). Impact of family
structure and quality on delinquency: A comparative assessment of structural and
functional factors. Criminology, 2_6, 235-261.

Wampold, B. E., & Freund, R D. (1987). Use of multiple regression in counseling psychology
research: A flexible data-analytc strategy. J ourna of Counselr_ng' chhology, 84, 372-
382.

Ward, A J. (1966). The meaning of movement responses and of its changes during therapy:
A review. JoumA of Pesonality Assesament. E, 418-428.

Weber, C. (1990). Aaahgsis of attachment through texture, human movement and human
content Rorschach variables in inpatient conduct disordered and dysthymic adolescents

Unpublished doctoral dissertation, United States Internatiorml University, San Diego,
CA

106

Weber, A, Meloy, J. R, & Gacono, C. B. (1992). A Rorschach study of attachment and
anxiety in inpatient conduct-disordered & dysthemic adolescents. Journal of
Persoflg Assesament. §§, 16-26.

 

Wells, L. E., & Rankin, J. H. (1988). Direct parental controls and delinquency. Criminology,
E, 263-285.

VVrdom, C. S. (1989a). Child abuse, neglect, and adult behavior: Research design and findings
on criminality, violence and child abuse. American J omaal of Orthopsychia_try, 2,
355-367.

Wrdom, C. S. (1989b). The cycle of violence. Science. 244, 160-166.

 

erlock, B. (1986). Narcissistic vulnerability in the hyperagressive child: The disregarded
(unloved, uncared-for) self. Pyohoanalytic Psychology, 8, 59-80.

 

Willock, B. (1987). The devalued (unloved, repugnant) self: A second facet of narcissistic
vulnerability in the aggressive, conduct-disordered child. chhoanalytic Psychology,
4, 219-240.

Wmnicott, D. W. (1958). Collected papers. New York: Basic Books.

Winnicott, D. W. (1965). The maturationalanocesses and the facilitating environment. New
York: International Universities Press, Inc.

Wood, 1. M., Nezworski, M. T., & Stejskal, W. J. (1996). The comprehensive system for the
Rorschach: A critical examination. chhological Science, 1, 3-10.

Zager, R, Arbit, J ., Hughes, J. R, Bussell, R E., & Busch, K. (1989). Developmental and
dismptive behavior disorders among delinquents. J oumal of the Ameriacaan Academy of
Child and Adolescent Psychiag, 3, 437-440.

'l'lll‘linll III!~'.I'|II|II .rll ‘I‘IDIIII ‘lll, .I ...l I. II11I'I' .

. ‘r ‘1 I ' lialll “ I III. I