Edited by
Eric J. Mash
and
Russell A. Barkley
THE GUILFORD PRESS
New York London
© 2003 The Guilford Press
A Division of Guilford Publications, Inc.
72 Spring Street, New York, NY 10012
www.guilford.com
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical,
photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher.
Printed in the United States of America
This book is printed on acid-free paper.
Last digit is print number: 9 8 7 6 5 4 3 2 1
Library of Congress Cataloging-in-Publication Data
Child psychopathology / edited by Eric J. Mash and Russell A. Barkley.
— 2nd ed.
p. cm.
Includes bibliographical references and index.
ISBN 1-57230-609-2
1. Child psychopathology. I. Mash, Eric J. II. Barkley, Russell A.,
1949– .
RJ499 .C4863 2002
618.92'89—dc21 2002009086
To our wives, Heather and Pat
This page intentionally left blank
About the Editors
Eric J. Mash, PhD, is Professor in the Department of Psychology at the University of Calgary. He
completed his undergraduate studies at City University of New York, his doctorate in clinical
psychology at Florida State University, and his postdoctoral work at the Oregon Health Sciences
University. Dr. Mash is a fellow of the American and Canadian Psychological Associations and has
served as an editorial board member and consultant for numerous scientific and professional
journals. His research interests are in child and adolescent psychopathology, assessment, and
therapy, and he has published many books and journal articles on these topics. His research has
focused on interaction patterns in families of children with different problems including attentiondeficit
and oppositional disorders and children who have been maltreated.
Russell A. Barkley, PhD, is Professor in the College of Health Professions at the Medical
University of South Carolina, Charleston, South Carolina. He is a Diplomate in both Clinical
Psychology and Clinical Neuropsychology, has written more than 200 scientific articles and book
chapters dealing with ADHD and related topics, and is author, editor, or coeditor of
15 books. Dr. Barkley is the founding Editor of The ADHD Report, a newsletter for clinicians,
and creator of seven professional videos, two of which have won national awards. He has served
as President of the International Society for Research in Child and Adolescent Psychopathology
and the Section of Clinical Child Psychology of the American Psychological Association (now
Division 53).
vii
This page intentionally left blank
Contributors
Anne Marie Albano, PhD, Institute for the Study
of Child and Adolescent Anxiety Disorders, New
York University Child Study Center, New York,
New York
Joan Rosenbaum Asarnow, PhD, Department
of Psychiatry, Neuropsychiatric Institute, University
of California, Los Angeles, School of Medicine,
Los Angeles, California
Robert F. Asarnow, PhD, Department of Psychiatry,
Neuropsychiatric Institute, University of
California, Los Angeles, School of Medicine, Los
Angeles, California
Russell A. Barkley, PhD, College of Health
Professions, Medical University of South Carolina,
Charleston, South Carolina
David H. Barlow, PhD, Center for Anxiety and
Related Disorders, Department of Psychology,
Boston University, Boston, Massachusetts
Marcia C. Barnes, PhD, Department of Pediatrics,
University of Toronto and The Hospital for
Sick Children, Toronto, Ontario, Canada
Carolyn Black Becker, PhD, Graduate School
of Professional and Applied Psychology, Rutgers
University, Piscataway, New Jersey
ix
Diane Benoit, MD, FRCPC, Department of
Psychiatry, University of Toronto and The Hospital
for Sick Children, Toronto, Ontario, Canada
Kim B. Burgess, PhD, Department of Human
Development, University of Maryland, College
Park, Maryland
Laurie Chassin, PhD, Department of Psychology,
Arizona State University, Tempe, Arizona
Bruce F. Chorpita, PhD, Department of Psychology,
University of Hawaii, Honolulu, Hawaii
Geraldine Dawson, PhD, Department of
Psychology, University of Washington, Seattle,
Washington
David J. A. Dozois, PhD, Department of Psychology,
University of Western Ontario, London,
Ontario, Canada
Elisabeth M. Dykens, PhD, Neuropsychiatric
Institute, Department of Child Psychiatry, University
of California, Los Angeles, Los Angeles,
California
Jack M. Fletcher, PhD, Department of Pediatrics/
Center for Academic and Reading Skills,
University of Texas–Houston Health Science
Center, Houston, Texas
Kenneth E. Fletcher, PhD, University of
Massachusetts Medical School, Worcester,
Massachusetts
Shelly Grabe, MA, Department of Psychological
Sciences, University of Missouri–Columbia,
Columbia, Missouri
Constance Hammen, PhD, Department
of Psychology, University of California, Los
Angeles, Los Angeles, California
Karen Heffernan, PhD, private practice, New
York, New York
Stephen P. Hinshaw, PhD, Department of
Psychology, University of California, Berkeley,
Berkeley, California
Robert M. Hodapp, PhD, Department of Education,
University of California, Los Angeles, Los
Angeles, California
Amy E. Kennedy, BA, Department of Human
Development, University of Maryland, College
Park, Maryland
Kevin M. King, MA, Department of Psychology,
Arizona State University, Tempe, Arizona
Laura Grofer Klinger, PhD, Department of
Psychology, University of Alabama, Tuscaloosa,
Alabama
Steve S. Lee, MA, Department of Psychology,
University of California, Berkeley, Berkeley,
California
G. Reid Lyon, Child Development and Behavior
Branch, National Institute of Child Health and
Human Development, Bethesda, Maryland
Karlen Lyons-Ruth, PhD, Department of Psychiatry,
Harvard Medical School, Cambridge,
Massachusetts
Eric J. Mash, PhD, Department of Psychology,
University of Calgary, Calgary, Alberta, Canada
x Contributors
Lizette Peterson, PhD, (deceased), Department
of Psychological Sciences, University of
Missouri–Columbia, Columbia, Missouri
Kelle Reach, BA, Department of Psychological
Sciences, University of Missouri–Columbia,
Columbia, Missouri
Peggy Renner, PhD, Department of Psychology,
University of Alabama, Tuscaloosa, Alabama
Jennifer Ritter, MA, Department of Psychology,
Arizona State University, Tempe, Arizona
Kenneth H. Rubin, PhD, Department of
Human Development, University of Maryland,
College Park, Maryland
Karen D. Rudolph, PhD, Department of
Psychology, University of Illinois, Champaign,
Illinois
Shannon L. Stewart, PhD, Child and Parent
Resource Institute, Ministry of Community, Family,
and Children’s Services, London, Ontario,
Canada
Ryan S. Trim, MA, Department of Psychology,
Arizona State University, Tempe, Arizona
Christine Wekerle, PhD, Department of Psychiatry,
University of Toronto and Child Psychiatry
Program, Centre for Addiction and Mental
Health, Toronto, Ontario, Canada
G. Terence Wilson, PhD, Graduate School of
Applied and Professional Psychology, Rutgers
University, Piscataway, New Jersey
David A. Wolfe, PhD, Department of Psychology,
The University of Western Ontario, London,
Ontario, Canada
Charles H. Zeanah, MD, Department of Psychiatry
and Pediatrics, Tulane University School
of Medicine, New Orleans, Louisiana
xi
Preface
Research in child, adolescent, and developmental
psychopathology continues to flourish,
even more so than when the first edition of this
text was published. Previously recognized disorders
are even better delineated than they were
only a few years ago, and a few new ones seem to
have been discovered along the way. The publication
rate in this field is extraordinary, with many
journals now focusing exclusively on childhood
mental illness and health, and numerous articles
on children appearing each month in journals that
were once the exclusive domains of adult psychopathology.
To those of us who take a developmental
view of psychopathology, this is a most gratifying
state of affairs as we come to recognize the
roots of many adult disorders in childhood and
adolescence. The down side, of course, is that
even the expert researchers in the various disorders
that constitute this field find it harder than
ever to keep abreast of research findings appearing
at such a rapid clip. And woe to the clinical
professionals who must deal with these childhood
disorders: They may find themselves quickly and
hopelessly behind in the advancements occurring
in the understanding of these clinical conditions.
Hence the need for a volume such as this, and
especially for its second edition, to assist the clinical
professional, student, and even expert in remaining
current on child and adolescent psychopathological
disorders.
Now more than ever, the field of child psychopathology
epitomizes the dynamic, accumulative,
and self-correcting nature of the scientific
enterprise, as new findings expand upon and are
assimilated with the established facts in any
given disorder. Often these new findings challenge
older theoretical or conceptual assumptions
or more explicit models of these disorders,
at times even leading to small-scale paradigm
shifts in perspective. In short, the literature on
child and adolescent psychopathology is alive,
well, prosperous, and rapidly advancing. Old
questions undoubtedly get answered, but along
the way those answers raise new questions for
researchers to pursue in ever more complex programs
of research on each of the childhood disorders
covered here. Although the pace and excitement
levels vary considerably across different
areas of child psychopathology, within each area
the eager anticipation of new knowledge remains
palpable as new lines of research and methodologies—
such as neuroimaging, behavioral and
molecular genetics, structural equation modeling,
and longitudinal designs—come to overlap old
ones and so provide greater opportunities to better
understand these disorders.
The challenge remains for this second edition
as it was for the first: How are we to capture the
current status of this rapidly evolving field? Our
answer was again to identify those experts who
have dedicated their professional careers to these
disorders, and let them—unfettered by fashion
or the editors’ pet perspectives—tell us what they
have learned. In other words, we tried to find the
most knowledgeable professionals on particular
disorders and asked them to provide up-to-date
and comprehensive summaries of the nature of
the disorders in which they have specialized. We
asked only that their discussions be grounded in
their respective bodies of scientific literature, eschewing
clinical lore, dogmatic wisdom, the sayings
of the guru du jour, or political agendas. We
also asked that they set aside the concerns of assessment
and treatment of their respective disorders,
so as to have ample room for the burgeoning
findings on the disorders themselves. These
other topics are the focus of related books (Mash
& Barkley, 1998; Mash & Terdal, 1997).
In essence, each author or group of authors was
once more challenged to answer these basic questions:
“What do we know about this disorder?”,
“What are the implications for future research
into further understanding the disorder?”, and,
just as important, “Where are the current limitations
or gaps in our knowledge that deserve future
attention?” If sound, scientifically grounded
theoretical or conceptual models of the disorder
exist, then these were also to be reviewed. In addressing
these questions, the experts assembled
here were directed to cover (1) the nature of the
behavior, symptoms, and/or cognitive and emotional
deficits that typify the core of each disorder;
(2) a brief historical perspective; (3) any criteria
that exist to establish its presence (diagnosis)
and a candid appraisal of those criteria; (4) epidemiological
knowledge pertaining to the prevalence,
gender distribution, and ethnic and cultural
factors associated with the disorder; (5) the
developmental course and varied pathways shown
xii Preface
to be associated with the disorder; (6) the psychiatric,
psychological, and social disorders or difficulties
that most often coexist with the disorder
(comorbidity); and (7) a survey of those things
believed to give rise to the disorder (etiology).
Once more, we believe that the many authors
assembled here have done a marvelous job accomplishing
their charge. We trust the reader will
concur.
As before, we are indebted to the professionals
who agreed to write for this second edition on
their respective disorders. We genuinely appreciate
the substantial time commitment they have
made to writing their chapters, many of which
are major updates of their previous work. Many
others deserve our gratitude as well, including
Jeannie Tang, Judith Grauman, Marie Sprayberry,
Carolyn Graham, Kim Miller, and Alison Wiigs,
for shepherding the manuscript through the
production process. Special thanks are also owed
to our long-time friends and founders of The
Guilford Press, Seymour Weingarten (Editor-in-
Chief) and Bob Matloff (President), for more
than 20 years of support for our various books,
including this one. Last, but hardly least, we thank
our families—Heather Mash, and Pat, Ken, and
Steve Barkley—for relinquishing the family time
such a project requires, and for their support, patience,
and encouragement of our careers in this
field.
Eric J. Mash, PhD
Department of Psychology
University of Calgary
Russell A. Barkley, PhD
College of Health Professions
Medical University of South Carolina
REFERENCES
Mash, E. J., & Barkley, R. A. (Eds.). (1998). Treatment of
childhood disorders (2nd ed.). New York: Guilford Press.
Mash, E. J., & Terdal, L. G. (Eds.). (1997). Assessment of
childhood disorders (3rd ed.). New York: Guilford Press.
Contents
xiii
III. INTRODUCTION
ONE Child Psychopathology: A Developmental– 3
Systems Perspective
Eric J. Mash and David J. A. Dozois
III. BEHAVIOR DISORDERS
TWO Attention-Deficit/Hyperactivity Disorder 75
Russell A. Barkley
THREE Conduct and Oppositional Defiant Disorders 144
Stephen P. Hinshaw and Steve S. Lee
FOUR Adolescent Substance Use Disorders 199
Laurie Chassin, Jennifer Ritter, Ryan S. Trim,
and Kevin M. King
III. EMOTIONAL AND SOCIAL DISORDERS
FIVE Childhood Mood Disorders 233
Constance Hammen and Karen D. Rudolph
SIX Childhood Anxiety Disorders 279
Anne Marie Albano, Bruce F. Chorpita,
and David H. Barlow
SEVEN Childhood Posttraumatic Stress Disorder 330
Kenneth E. Fletcher
EIGHT Social Withdrawal in Childhood 372
Kenneth H. Rubin, Kim B. Burgess, Amy E. Kennedy,
and Shannon L. Stewart
IV. DEVELOPMENTAL AND LEARNING DISORDERS
NINE Autistic Disorder 409
Laura Grofer Klinger, Geraldine Dawson,
and Peggy Renner
TEN Childhood-Onset Schizophrenia 455
Joan Rosenbaum Asarnow and Robert F. Asarnow
ELEVEN Mental Retardation (Intellectual Disabilities) 486
Robert M. Hodapp and Elisabeth M. Dykens
TWELVE Learning Disabilities 520
G. Reid Lyon, Jack M. Fletcher, and Marcia C. Barnes
IV. INFANTS AND CHILDREN AT RISK FOR DISORDER
THIRTEEN Disorder and Risk for Disorder during Infancy 589
and Toddlerhood
Karlen Lyons-Ruth, Charles H. Zeanah,
and Diane Benoit
FOURTEEN Child Maltreatment 632
Christine Wekerle and David A. Wolfe
VI. EATING AND HEALTH-RELATED DISORDERS
FIFTEEN Eating Disorders 687
G. Terence Wilson, Carolyn Black Becker,
and Karen Heffernan
SIXTEEN Health-Related Disorders 716
Lizette Peterson, Kelle Reach, and Shelly Grabe
Author Index 751
Subject Index 786
xiv Contents
1. A Developmental–Systems Perspective 1
I
INTRODUCTION
2 I. INTRODUCTION
This page intentionally left blank
1. A Developmental–Systems Perspective 3
CHAPTER
ONE
Child Psychopathology
A Developmental–
Systems Perspective
Eric J. Mash
David J. A. Dozois
3
This volume provides a comprehensive account
of the characteristics, definitions, developmental
course, correlates, causes, contexts, and outcomes
of psychopathology in children.1 Our knowledge
base of child and developmental psychopathology
has grown exponentially over the past decade
(Cicchetti & Cohen, 1995a, 1995b; Cicchetti &
Sroufe, 2000; Mash & Wolfe, 2002; Ollendick &
Hersen, 1998). New conceptual frameworks, research
methods, and findings continue to advance
our understanding of childhood disorders
(Cicchetti & Rogosch, 1999; Rutter & Sroufe,
2000; Sameroff, Lewis, & Miller, 2000), as well
as our ability to assess and treat children with
problems (Mash & Barkley, 1998; Mash &
Terdal, 1997a; Orvaschel, Faust, & Hersen, 2001;
Shaffer, Lucas, & Richters, 1999). However, this
knowledge base is compromised by the frequently
atheoretical, unsystematic, and fragmented
fashion in which research findings in
child psychopathology have accrued, and by the
conceptual and research complexities inherent in
the study of such a rapidly changing and socially
embedded organism as the child (Hinshaw, 2001;
Jensen et al., 1993; Kazdin & Kagan, 1994). In
this introductory chapter, we address several central
themes and issues related to conceptualizing
childhood dysfunction and its many determinants.
In doing so, we provide a developmental–
systems framework for understanding child psychopathology—
one that emphasizes the role of
developmental processes, the importance of context,
and the influence of multiple and interacting
events and processes in shaping adaptive and
maladaptive development.
FACTORS COMPLICATING
THE STUDY OF CHILD
PSYCHOPATHOLOGY
Almost since modern views of mental illness began
to emerge in the late 18th and early 19th
centuries, far less attention has been given to the
study of psychopathology in children than in
adults (Silk, Nath, Siegel, & Kendall, 2000). For
example, in 1812 Benjamin Rush, the first American
psychiatrist, suggested that children were less
likely to suffer from mental illness than adults,
because the immaturity of their developing brains
would prevent them from retaining the mental
events that caused insanity (Silk et al., 2000).
More recently, interest in the study of child
psychopathology has increased dramatically. This
is due to a growing realization that (1) many childhood
problems have lifelong consequences and
costs both for children and for society; (2) most
adult disorders are rooted in early childhood con4
I. INTRODUCTION
ditions and/or experiences; and (3) a better
understanding of childhood disorders offers
promise for developing effective intervention and
prevention programs (National Advisory Mental
Health Council [NAMHC] Workgroup, 2001).
Issues concerning the conceptualization and
definition of psychopathology in children continue
to be vigorously debated. Such debates are
fueled by the relative absence of well-controlled
research studies with children as compared with
adults. Until recently, much of the field’s accumulated
knowledge about child psychopathology,
its causes, and its outcomes was extrapolated from
theory and research on adult disorders. For
example, only in the last 5–10 years have childfocused
models and research into such disorders
as depression and anxiety emerged (Zahn-Waxler,
Klimes-Dougan, & Slattery, 2000).
Even in studies conducted with children, much
of our knowledge is based on findings obtained
at a single point in a child’s development and in
a single context. Although useful, such findings
provide still photographs of moving targets and
fail to capture the dynamic changes over time that
characterize most forms of child psychopathology
(Achenbach & Dumenci, 2001; Lewis & Granic,
2000; Patterson, 1993). In addition, prior studies
have not given sufficient attention to the social
and cultural milieu in which atypical child
development occurs (Cicchetti & Aber, 1998;
Garc?a Coll & Garrido, 2000). Contextual models
(e.g., Bronfenbrenner, 1977) and longitudinal
approaches (e.g., Robins, 1966) have been
available in the field of child study for some time.
However, only in the past decade has the research
enterprise taken seriously the need for developmentally
sensitive systems-oriented models to
account for the emergence of psychopathology
in children (del Carmen & Huffman, 1996;
Sameroff, 2000a), or the need to study developmental
trajectories utilizing longitudinal methods
(e.g., Emery, Waldron, Kitzmann, & Aaron, 1999;
Hauser-Cram, Warfield, Shonkoff, & Krauss,
2001; Kotler, Cohen, Davies, Pine, & Walsh,
2001; Maughan & Rutter, 2001; Verhulst & Koot,
1991).
The study of child psychopathology is further
complicated by the facts that childhood problems
do not come in neat packages, and that most
forms of psychopathology in children are known
to overlap and/or to coexist with other disorders
(Angold, Costello, & Erkanli, 1999). For example,
there is much overlap among such problems
as violence, emotional and behavioral disorders,
child maltreatment, substance abuse, delinquency,
and learning difficulties (e.g., Greenbaum, Prange,
Friedman, & Silver, 1991); between childhood
anxiety and depression (e.g., Compas & Oppedisano,
2000; Seligman & Ollendick, 1998); and
between reading disabilities and anxiety and depression
(Willcutt & Pennington, 2000b). Many
behavioral and emotional disturbances in children
are also associated with specific physical
symptoms and/or medical conditions (Egger,
Costello, Erkanli, & Angold, 1999; Meltzer,
Gatward, Goodman, & Ford, 2000).
It is also the case that distinct boundaries between
many commonly occurring childhood difficulties
(e.g., noncompliance, defiance) and
those problems that come to be labeled as “disorders”
(e.g., oppositional defiant disorder) are
not easily drawn (e.g., Loeber, Burke, Lahey,
Winters, & Zera, 2000). Judgments of deviancy
often depend as much on other child characteristics
(e.g., age, sex, intelligence), the situational
appropriateness of a child’s behavior, the social
and cultural context in which judgments are
made, and the characteristics and decision rules
of adults who make these judgments as they do
on any specific behaviors displayed by the child
(Achenbach, 2000; Mash & Terdal, 1997b).
There is a growing recognition that all current
diagnostic categories of child psychopathology
are heterogeneous with respect to etiology and
outcome, and will need to be broken down into
subtypes (Kagan, 1997). Although these diagnostic
systems make some allowances for subtypes,
designations are rudimentary at best, given
the many different subgroups and types that
have been identified for children with such disorders
as attention-deficit/ hyperactivity disorder
(ADHD), conduct disorder, oppositional defiant
disorder, anxiety disorders, and mood disorders
(e.g., Milich, Balentine, & Lynam, 2001).
It has become increasingly evident that most
forms of child psychopathology cannot be attributed
to a single unitary cause. Although certain
rare disorders (e.g., phenylketonuria, fragile-X
mental retardation, or Rett’s disorder) may be
caused by single genes, current models in behavioral
and molecular genetics recognize that more
common and complex disorders are the result of
the operation of multigene systems containing
varying effect sizes (Goldsmith, Gottesman, &
Lemery, 1997; McGuffin, Riley, & Plomin, 2001;
O’Conner & Plomin, 2000). Most forms of child
psychopathology are polygenic, involving a number
of susceptibility genes that interact with one
1. A Developmental–Systems Perspective 5
another and with environmental influences to
result in observed levels of impairment (Rutter,
2000a; State, Lombroso, Pauls, & Leckman,
2000). Child and family disturbances are likely to
result from multiple, frequently co-occurring,
reciprocal, and interacting risk factors, causal
events, and processes (e.g., Eaves et al., 1997; Ge,
Conger, Lorenz, Shanahan, & Elder, 1995; Rende,
1999; Rutter et al., 1997). Contextual events exert
considerable influence in producing child and
adolescent disorders—an influence that is almost
always equivalent to or greater than those factors
usually thought of as residing “within” the child
(Caspi, Taylor, Moffitt, & Plomin, 2000; Reiss &
Neiderhiser, 2000; Rutter, 2000b).
Numerous determinants of child psychopathology
have been identified, including genetic
influences (e.g., State et al., 2000); hypo- or hyperreactive
early infant dispositions (e.g., Hirshfeld,
Biederman, Brody, & Faraone, 1997); insecure
child–parent attachments (e.g., Bretherton, 1995;
Sroufe, Carlson, Levy, & Egeland, 1999); difficult
child behavior (e.g., Costello & Angold,
2001); social-cognitive deficits (e.g., Crick &
Dodge, 1994; Schwartz & Proctor, 2000); deficits
in social learning (e.g., Patterson, 1982;
Patterson, Reid, & Dishion, 1992), emotion regulation
(e.g., Keenan, 2000), and/or impulse control
and response inhibition (Barkley, 1997; Nigg,
2000, 2001); neuropsychological and/or neurobiological
dysfunction (e.g., Cicchetti & Cannon,
1999); maladaptive patterns of parenting (e.g.,
Lovejoy, Graczyk, O’Hare, & Neuman, 2000);
parental psychopathology, such as maternal depressed
mood (e.g., Goodman & Gotlib, 1999);
parental or couple discord (e.g., Grych &
Fincham, 2001); limited family resources and
other poverty-related life stressors (e.g., Rutter,
1999); institutional deprivation (e.g., Kreppner
et al., 2001); and a host of other potential factors.
These factors cannot be understood in isolation,
and for most disorders, research does not support
granting central etiological status to any single risk
or causal factor (e.g., Seifer, Sameroff, Baldwin,
& Baldwin, 1992).
Since the many causes and outcomes of child
psychopathology operate in dynamic and interactive
ways over time, they are not easy to disentangle.
The designation of a specific factor as a
cause or an outcome of child psychopathology
usually reflects (1) the point in an ongoing developmental
process at which the child is observed,
and (2) the perspective of the observer. For example,
a language deficit may be viewed as a disorder
in its own right (e.g., mixed receptive–expressive
language disorder), the cause of other
difficulties (e.g., impulsivity), or the outcome of
some other condition or disorder (e.g., autistic
disorder). In addition, biological and environmental
determinants interact at all periods of
development. Dawson, Hessl, and Frey (1994),
for example, noted that the characteristic styles
parents use in responding to their infants’ emotional
expressions may influence how patterns of
cortical mappings and connections within the limbic
system are established in the infants. Similarly,
J. Hart, Gunnar, and Cicchetti (1995) reported
that maltreated preschoolers showed reduced
cortisol activity in response to stress relative to
controls—a finding that suggests altered activity
of the stress-regulating hypothalamic–pituitary–
adrenocortical (HPA) system among children
who have been maltreated. These and other findings
suggest that early experiences may shape
neural structure and function, which may then
create dispositions that direct and shape a child’s
later experiences and behavior (Cicchetti &
Walker, 2001; Dawson et al., 1999; Glaser, 2000;
Kaufman & Charney, 2001; Post & Weiss, 1997).
As will be discussed throughout this volume,
current models of child psychopathology seek
to incorporate the role of evolved mechanisms,
neurobiological factors, early parent–child relationships,
attachment processes, a long-term
memory store that develops with age and experience,
micro- and macrosocial influences, cultural
factors, age and gender, and reactions from the
social environment as variables and processes that
interact and transform one another over time. In
short, then, current approaches view the roots of
developmental and psychological disturbances in
children as the result of complex interactions over
the course of development between the biology
of brain maturation and the multidimensional
nature of experience (Cicchetti & Toth, 1997;
Cicchetti & Tucker, 1994; Reiss & Neiderhiser,
2000; Rutter et al., 1997).
The experience and the expression of psychopathology
in children are known to have cognitive,
affective, physiological, and behavioral
components; in light of this, many differing
descriptions and definitions of dysfunctionality in
children have been proposed. As we discuss in a
later section, a common theme in defining child
psychopathology has been that of “adaptational
failure” in one or more of these components or
in the ways in which these components are organized
and integrated (Sameroff, 2000a; Sroufe,
6 I. INTRODUCTION
1997; Sroufe & Rutter, 1984). Adaptational failure
may involve deviation from age-appropriate
norms (Achenbach, 2001), exaggeration or diminishment
of normal developmental expressions,
interference in normal developmental progress,
failure to master developmental tasks, failure to
develop a specific function or regulatory mechanism,
and/or the use of non-normative skills (e.g.,
rituals, dissociation) as a way of adapting to
regulatory problems or traumatic experiences
(Fischer et al., 1997; Sroufe, 1997).
A multitude of etiological models and treatment
approaches have been proposed to explain
and remediate psychopathology in children. Unfortunately,
most of these have yet to be substantiated—
or, for that matter, even tested (Kazdin,
2000, 2001). These models and approaches have
differed in their relative emphasis on certain
causal mechanisms and constructs, often using
very different terminology and concepts to describe
seemingly similar child characteristics and
behaviors. Although useful, many of these models
have been based on what seem to be faulty premises
concerning singular pathways of causal
influence that do not capture the complexities
of child psychopathology (Kazdin & Kagan,
1994).
In this regard, evolutionary models have emphasized
the role of selection pressures operating
on the human species over millions of years;
biological paradigms have emphasized genetic
mutations, neuroanatomy, and neurobiological
mechanisms as factors contributing to psychopathology;
psychodynamic models have focused on
intrapsychic mechanisms, conflicts, and defenses;
attachment models have emphasized the importance
of early relationships and the ways in which
internal representations of these relationships
provide the foundation for constructing working
models of self, others, and relationships more
generally; behavioral/reinforcement models have
emphasized excessive, inadequate, or maladaptive
reinforcement and/or learning histories;
social learning models have emphasized the importance
of observational learning, vicarious
experience, and reciprocal social interactions;
cognitive models generally focus on the child’s
distorted or deficient cognitive structures and
processes; affective models have emphasized
dysfunctional emotion-regulating mechanisms;
and family systems models have conceptualized
child psychopathology within a framework of
intra- and intergenerational family systems and
subsystems and have emphasized the structural
and/or functional elements that surround family
relational difficulties.
The distinctiveness of each model mentioned
above is in the relative importance it attaches to
certain events and processes. However, it should
be emphasized that despite these variations in the
relative emphasis given to certain causes versus
others, most models recognize the role of multiple
interacting influences. For example, although
differing in emphasis, social learning and affective
models both place importance on the role of
symbolic representational processes in explaining
childhood dysfunction.
There is a growing recognition of the need to
integrate currently available models through
intra- and interdisciplinary research efforts. Such
integration generally requires looking beyond the
emphasis of each single-cause theory to see what
can be learned from other approaches, as well as
a general openness to relating concepts and findings
from diverse theories (cf. Arkowitz, 1992).
Recent studies suggest that theoretical integration
is becoming more common in psychopathology
research (e.g., Beauchaine, 2001).
Attachment theory has, for instance, been increasingly
integrated with cognitive models (e.g.,
Ingram & Ritter, 2000). Theoretical integration
is also apparent in studies combining proximal
cognitive and interpersonal factors with distal
variables, such as the early home environment
and patterns of attachment (e.g., Lara, Klein, &
Kasch, 2000). The link between cognitive and
neuropsychological functioning is likewise being
tested more frequently (e.g., Nigg, Blaskey,
Huang-Pollack, & Rappley, 2002; Pine & Grun,
1999). Thus it appears as though researchers are
beginning to recognize the importance of combining
theoretical approaches, and are accepting
the monumental task of incorporating increased
complexity into their research designs.
Interdisciplinary perspectives on child psychopathology
mirror the considerable investment in
children on the part of many different disciplines
and professions. The study of the etiology and
maintenance of psychopathology in children has
been and continues to be the subject matter of
psychology, medicine, psychiatry, education, and
numerous other disciplines. Clearly, no one discipline
has proprietary rights to the study of childhood
disturbances. Each discipline has formulated
child psychopathology in terms of its own
unique perspective. Particularly relevant, in the
context of this chapter, is that child psychopathology
and normality in medicine and psychiatry
1. A Developmental–Systems Perspective 7
have typically been conceptualized and defined
categorically in terms of the presence or absence
of a particular disorder or syndrome that is believed
to exist “within the child.” In contrast, psychology
has more often conceptualized psychopathology–
normality as representing extremes on
a continuum or dimension of characteristics, and
has also focused on the role of environmental
influences that operate “outside the child.” However,
the boundaries between categories and
dimensions, or between inner and outer conditions
and causes, are arbitrarily drawn, and
there is a continuing need to find workable ways
of integrating the two different world views of
psychiatry/medicine and psychology (Richters &
Cicchetti, 1993; Scotti & Morris, 2000; Shaffer
et al., 1999).
As the subsequent chapters in this volume attest,
research into child psychopathology is accelerating
at a remarkable rate. This in turn has
resulted in a rapidly expanding and changing
knowledge base. Each chapter in this volume
provides a comprehensive review of current research
and theory for a specific form of child psychopathology,
and a discussion of new developments
and directions related to this disorder. In
the remainder of this introductory chapter, we
provide a brief overview and discussion of the
following: historical developments in the study of
child psychopathology; epidemiological considerations;
basic issues; approaches to the definition
and classification of childhood disorders; common
types of psychopathology in children; important
philosophical and epistemological assumptions
that have guided theory and research;
predominant theories regarding etiology; and
prevalent and recurrent conceptual and methodological
issues that cut across the wide spectrum
of disorders represented in this volume. Particular
emphasis is given to concepts, methods, and
strategies capturing the complexities, reciprocal
influences, and divergent pathways that current
models and research have identified as crucial for
understanding child psychopathology.
HISTORICAL CONTEXT FOR CHILD
PSYCHOPATHOLOGY
Brief Historical Overview
Historical developments surrounding the emergence
of child psychopathology as a field of study
have been documented in a number of excellent
sources and are considered only briefly here (see
Achenbach, 1982; Cicchetti, 1990; Donohue,
Hersen, & Ammerman, 2000; Kanner, 1962; Rie,
1971; Rubinstein, 1948; Silk et al., 2000). In general,
the emergence of concepts of child psychopathology
was inextricably related to the broader
philosophical and societal changes in the ways
children have been viewed and treated by adults
over the course of history (Aries, 1962; Borstelmann,
1983; French, 1977; Postman, 1994). Several
overlapping perspectives for conceptualizing
and dealing with deviant child behavior emerged,
including the religious, the legal, the medical, the
social, and the educational (Costello & Angold,
2001).
In ancient Greek and Roman societies, child
behavior disorders were believed to result from
organic imbalances, and children with physical
or mental handicaps, disabilities, or deformities
were viewed as sources of economic burden and/
or social embarrassment. As such, they were usually
scorned, abandoned, or put to death (French,
1977). This mistreatment, by today’s standards,
was common throughout the Middle Ages (A.D.
500–1300). In colonial America, as many as twothirds
of all children died prior to the age of
5 years, and those who survived continued to be
subjected to harsh treatment by adults. For example,
the Massachusetts Stubborn Child Act of
1654 permitted a father to petition a magistrate
to put a “stubborn” or “rebellious” child to death
(fortunately, no sentences were carried out); in
Massachusetts and elsewhere, mentally ill children
were kept in cages and cellars into the mid-
1800s (Silk et al., 2000).
The historical record indicates that prior to
the 18th century, when references to disordered
child behavior were made at all, they were usually
presented in terms of the problem child’s
behavior as inherently evil (Kanner, 1962). Bizarre
behaviors in children were attributed to
Satanic possession and evil spirits during the
Spanish Inquisition, and both John Calvin and
Martin Luther viewed mentally retarded children
as filled with Satan. And, as noted by Rie (1971),
“No distinct concept of disordered behavior
in children could emerge so long as possession
by the devil excluded other notions of causality”
(p. 8).
Although nearly all varieties of aberrant behavior
in children have existed for millennia, the formal
study of such behavior is relatively recent.
Following a comprehensive review of historical
developments in child psychopathology, Rie
8 I. INTRODUCTION
(1971) concluded: “There is a consensus, then,
about the absence of any substantial body of
knowledge—prior to the twentieth century—
concerning disordered behavior in childhood;
about the inconsistencies and discontinuities of
efforts on behalf of disturbed children; and about
the relative absence of those professional specialties
which now concern themselves with such
problems” (p. 6). Rubinstein (1948) noted that (1)
there was not a single article dealing with insanity
in childhood in any of the first 45 volumes of
the Journal of Insanity; (2) there was no discovery
or theory of importance to child psychiatry in
the American literature prior to 1900, and no
research today stems from any of these writings;
and (3) the only significant work with children
prior to the 20th century focused on the care,
treatment, and training of “mental defectives.”
Increased concern for the plight and welfare
of children with mental and behavioral disturbances
was the result of two important influences.
First, advances in general medicine, physiology,
and neurology led to the reemergence of
the organic disease model and a concomitant
emphasis on more humane forms of treatment.
Second, the growing influence of the philosophies
of John Locke, Johann Pestalozzi, and
Jean-Jacques Rousseau led to the view that children
needed moral guidance and support. With
these changing views came an increased concern
for moral education, compulsory education, and
improved health practices. These early influences
also provided the foundation for evolving views
of child psychopathology as dependent on both
organic and environmental causes.
Masturbatory Insanity:
An Example
Societal and clinical views regarding masturbation
in children can be used to illustrate the ways in
which conceptualizations of child psychopathology
have changed over time, as well as several
general issues related to its definition, study, and
treatment. In addition to the historical significance
of masturbation as the first disorder described
as unique to children and adolescents
(Rie, 1971), early conceptualizations of masturbatory
insanity illustrate a view of mental illness
as residing within the child (Cattell, 1938; Hare,
1962; Rees, 1939; Rie, 1971; Szasz, 1970).
Society’s objections to masturbation originated
from Orthodox Jewish codes and from Judeo-
Christian dogmata (Patton, 1985; Szasz, 1970). It
was not until the 18th century—with a decline in
the domination of religious thought, coupled with
the augmented influence of science—that masturbation
came to be viewed as particularly
harmful (Rie, 1971; Szasz, 1970). An anonymous
clergyman who later became a physician wrote a
dissertation entitled Onania, or the Heinous Sin
of Self-Pollution (circa 1710, cited in Szasz, 1970).
It was this manuscript that initially transformed
the moral convictions regarding the wrongfulness
of masturbation into a physiological explanation
with severe medical ramifications. Following this
exposition, numerous books appeared claiming
that masturbation was a predominant etiological
cause of both physical disease and mental illness.
Thus the notion that sexual overindulgence was
deleterious to one’s health was accepted, “virtually
unaltered, first by the Church and then by
Medicine” (Szasz, 1970, p. 182). Although the
medical view of masturbation first emphasized
the adverse impact upon physical health, the
dominant thinking shifted by the middle of the
19th century to a focus on the presumed negative
effects on mental health and nervous system
functioning. By the latter part of the 19th century,
masturbation was the most frequently mentioned
“cause” of psychopathology in children. In fact,
Spitzka (1890; cited in Rie, 1971) attributed at
least 25% of all psychiatric cases to this etiological
factor.
Views of masturbatory insanity emerged and
were maintained in the absence of any thought
to the contrary, and without any consideration of
the base rate of masturbation in the general population.
Although interest in masturbatory insanity
began to wane in the latter half of the 19th
century, the argument endured (albeit in milder
forms) during the early 20th century, when
psychoanalytic theory gained rapid acceptance.
Freud suggested that masturbation was one of the
precipitants of neurasthenia, hypochondriasis,
and anxiety neurosis (Rees, 1939). Apart from his
own theories regarding the pathogenesis of neuroses,
however, Freud did not present any real
evidence for this view (Szasz, 1970). Eventually
the notion of masturbatory insanity gave way to
the concept of neurosis, but it was still not until
much later in the 20th century that the misguided
and illusory belief in a relationship between masturbation
and mental illness was dispelled.
As conceptualizations of childhood psychopathology
evolved, and several variants of psychotherapy
and residential treatments were developed
(see Grellong, 1987, and Roberts & Kurtz,
1. A Developmental–Systems Perspective 9
1987), the search for determinants of psychiatric
disorders in children became increasingly sophisticated,
thorough, and systematic (Rie, 1971).
With this increased refinement of theory and
research, there remained only fragments of the
etiological hypothesis of masturbation. For example,
in some psychoanalytic circles, enuresis
was thought to symbolize suppressed masturbation
(Rees, 1939; Walker, Kenning, & Faust-
Campanile, 1989). Eventually masturbation came
to be viewed as entirely harmless (Szasz, 1970)
and even as usefully adaptive (Baker, 1996).
This brief historical review illustrates a number
of points. First, it shows how the political and
social climates influence our definitions of child
psychopathology. The impact of religious thought
was clearly reflected in the transformation from
the moral judgment against the sins of the flesh,
to the medical opinion that masturbation was
harmful to one’s physical health, to the psychiatric
assertion that sexual overindulgence caused
insanity.
Second, the review points out the need to be
cognizant of the ways in which moral convictions,
idiosyncratic definitions of normality or pathology,
and personal expectations influence what
investigators look for and ultimately find in the
name of science. In the case of masturbation,
misleading findings resulted because hypotheses
were “tested” with a mentality of confirmation
rather than falsification (see Maxwell & Delaney,
1990). Szasz (1970), in writing about the powerful
authority of America’s historical psychiatric
figures such as Benjamin Rush, noted that there
is a tendency among scientists to “attend only to
those of their observations that confirm the accepted
theories of their age, and reject those that
refute them” (p. 187).
Third, masturbatory insanity illustrates the
potential dangers that ensue when treatment
decisions are made on the basis of deficient theoretical
exposition and in the absence of empirical
data. For example, early treatments consisted
of clitoridectomies for women and spike-toothed
rings placed on the penises of men (Szasz, 1970).
Finally, the example of masturbatory insanity
portrays the long-standing view of psychopathology
as residing within the child and the essential
neglect of the role of his or her surroundings,
context, relationships, and the interactions among
these variables.
Current theory, research, and practice reflect
a shift toward acknowledging developmental factors
and including the family, peer group, school,
and other sources of influence in conceptualizing
and understanding child psychopathology
(Luthar, Burack, Cicchetti, & Weisz, 1997; Mash
& Wolfe, 2002). Additional developments have
included an increased research emphasis on examining
the interactions of multiple proximal and
distal vulnerability factors (Ingram, Miranda, &
Segal, 1998; Price & Lento, 2001), understanding
psychopathology across the life span (Ingram
& Price, 2001), identifying empirically supported
treatments for various childhood problems
(Kazdin & Weisz, 1998; Lonigan, Elbert, &
Johnson, 1998), and a focus on prevention
(Greenberg, Domitrovich, & Bumbarger, 2001;
National Institute of Mental Health, 2001).
SIGNIFICANCE OF
CHILD PSYCHOPATHOLOGY
There has been and continues to be a great
deal of misinformation and folklore concerning
disorders of childhood. Many unsubstantiated
theories have existed in both the popular and scientific
literatures. These have ranged from mid-
19th-century views that overstimulation in the
classroom causes insanity (see Makari, 1993), to
mid-20th-century views that inadequate parenting
causes autism (Bettelheim, 1967) or that chemical
food additives cause hyperactivity (Feingold,
1975). In addition, many of the constructs used
to describe the characteristics and conditions of
psychopathology in children have been globally
and/or poorly defined (e.g., “adjustment problem,”
“emotional disturbance”). Despite the limitations,
uncertainties, and definitional ambiguities
that exist in the field, it is also evident that
psychopathology during childhood represents a
frequently occurring and significant societal concern
that is gradually coming to the forefront of
the political agenda.
In the United States, the approach of the new
millennium witnessed the first Surgeon General’s
report on mental health (U.S. Public Health Service,
1999), which was followed by White House
meetings on mental health in young people and
on the use of psychotropic medications with children.
A Surgeon General’s conference on children’s
mental health resulted in an extensive report
and recommendations (U.S. Public Health
Service, 2001a), a similar report on youth violence
(U.S. Public Health Service, 2001b), and a “blueprint”
for research on child and adolescent mental
health (NAMHC Workgroup, 2001).2
10 I. INTRODUCTION
Increasingly, researchers in the fields of developmental
psychopathology, child psychiatry, and
clinical child psychology are considering the social
policy implications of their work and striving
to effect improvements in the identification of
and services for youths with mental health needs
(Cicchetti & Toth, 2000; Weisz, 2000). Greater
recognition is also being given to factors that
contribute to children’s successful mental functioning,
personal well-being, productive activities,
fulfilling relationships, and ability to adapt to
change and cope with adversity (Cicchetti, Rappaport,
Sandler, & Weissberg, 2000; Thompson
& Ontai, 2000; U.S. Department of Health and
Human Services, 2000b; U.S. Public Health Service,
2001a).
The growing attention to children’s mental
health problems and competencies arises from a
number of sources. First, many young people
experience significant mental health problems
that interfere with normal development and functioning.
As many as 1 in 5 children in the United
States experience some type of difficulty (Costello
& Angold, 2000; Roberts, Attkisson, & Rosenblatt,
1998), and 1 in 10 have a diagnosable disorder
that causes some level of impairment
(Burns et al., 1995; Shaffer et al., 1996). These
numbers probably underestimate the magnitude
of the problem, since they do not include a substantial
number of children who manifest subclinical
or undiagnosed disturbances that may
place them at high risk for the later development
of more severe clinical problems. For example,
McDermott and Weiss (1995) reported that of
the children in their national sample who were
classified as adjusted, 34.4% were classified as
being only “marginally” adjusted. In addition,
although not meeting formal diagnostic criteria,
many subclinical conditions (e.g., depressed
mood, eating problems) are also associated with
significant impairment in functioning (e.g., Angold,
Costello, Farmer, Burns, & Erkanli, 1999; Lewinsohn,
Striegel-Moore, & Seeley, 2000). Evidence
gathered by the World Health Organization
(WHO) suggests that by the year 2020, childhood
neuropsychiatric disorders will rise by over 50%
internationally, to become one of the five most
common causes of morbidity, mortality, and disability
among children (U.S. Public Health Service,
2001a).
Second, a significant proportion of children do
not grow out of their childhood difficulties, although
the ways in which these difficulties are
expressed change in both form and severity over
time (Offord et al., 1992). Even when diagnosable
psychopathology is not evident at later ages,
a child’s failure to adjust during earlier developmental
periods may still have a lasting negative
impact on later family, occupational, and social
adjustment. And some forms of child psychopathology—
for example, an early onset of antisocial
patterns of behavior in boys—can be highly predictive
of various negative psychosocial, educational,
and health outcomes in adolescence and
adulthood (see Hinshaw & Lee, Chapter 3, this
volume).
Third, recent social changes and conditions
may place children at increasing risk for the development
of disorders, and also for the development
of more severe problems at younger
ages (Duncan, Brooks-Gunn, & Klebanov, 1994;
Kovacs, 1997). These social changes and conditions
include multigenerational adversity in inner
cities; chronic poverty in women and children;
pressures of family breakup, single parenting, and
homelessness; problems of the rural poor; direct
and indirect exposure to traumatic events (e.g.,
terrorist attacks or school shootings); adjustment
problems of children in immigrant families; difficulties
of Native American children; and conditions
associated with the impact of prematurity,
HIV, cocaine, and alcohol on children’s growth
and development (McCall & Groark, 2000; National
Commission on Children, 1991; Shonkoff
& Phillips, 2000). In addition to sociocultural
changes, medical advances associated with
higher rates of fetal survival may also contribute
to a greater number of children showing
serious behavior problems and learning disorders
at a younger age.
Fourth, for a majority of children who experience
mental health problems, these problems go
unidentified: Only about 20% receive help, a statistic
that has not changed for some time (Burns
et al., 1995). Even when children are identified
and receive help for their problems, this help may
be less than optimal. For example, only about half
of children with identified ADHD seen in realworld
practice settings receive care that conforms
to recommended treatment guidelines (Hoagwood,
Kelleher, Feil, & Comer, 2000). The fact
that so few children with mental health problems
receive appropriate help is probably related to
such factors as a lack of screening, inaccessibility,
cost, a lack of perceived need on the part of
parents, parental dissatisfaction with services, and
1. A Developmental–Systems Perspective 11
the stigmatization and exclusion often experienced
by these children and their families
(Hinshaw & Cicchetti, 2000; Kroes et al., 2001).
Empirically supported prevention and treatment
programs for many childhood disorders are only
now becoming available (Kazdin & Weisz, 1998;
Lonigan et al., 1998), and there is a pressing need
for the development and evaluation of prevention
and intervention programs that are grounded in
theory and research on child development in
general, and developmental psychopathology in
particular (Greenberg et al., 2001; Kazdin, 2001;
Kurtines & Silverman, 1999; NAMHC Workgroup,
2001; Rapport, 2001).3
Fifth, a majority of children with mental health
problems who go unidentified and unassisted
often end up in the criminal justice or mental
health systems as young adults (Loeber & Farrington,
2000). They are at much greater risk for
dropping out of school and of not being fully functional
members of society in adulthood; this adds
further to the costs of childhood disorders in
terms of human suffering and financial burdens.
For example, average costs of medical care for
youngsters with ADHD are estimated to be
double those for youngsters without ADHD
(Leibson, Katusic, Barberesi, Ransom, & O’Brien,
2001). Moreover, allowing just one youth to leave
high school for a life of crime and drug abuse is
estimated to cost society from $1.7 to $2.3 million
(Cohen, 1998).
Finally, a significant number of children in
North America are being subjected to maltreatment,
and chronic maltreatment during childhood
is associated with psychopathology in
children and later in adults (Emery & Laumann-
Billings, 1998; MacMillan et al., 2001). Based on
a review of the evidence, De Bellis (2001) has
proposed that the psychobiological outcomes of
abuse be viewed as “an environmentally induced
complex developmental disorder” (p. 539). Although
precise estimates of the rates of occurrence
of maltreatment are difficult to obtain, due
to the covert nature of the problem and other
sampling and reporting biases (see Cicchetti &
Manly, 2001; Mash & Wolfe, 1991), the numbers
appear to be large. Nearly 3 million suspected
cases of child abuse and neglect are investigated
each year by child protective service agencies, and
about 1 million children in the United States were
confirmed as victims of child maltreatment in
1998 (U.S. Department of Health and Human
Services, 2000a). It has been estimated that each
year as many as 2,000 infants and young children
die from abuse or neglect at the hands of their
parents or caregivers (U.S. Advisory Board on
Child Abuse and Neglect, 1995). Moreover, many
reports of “accidental” injuries in children may be
the result of unreported mistreatment by parents
or siblings (Peterson & Brown, 1994). It would
appear, then, that the total number of children
who show adverse psychological and physical effects
of maltreatment in North American society
is staggering.
EPIDEMIOLOGICAL
CONSIDERATIONS
Prevalence
Epidemiological studies seek to determine the
prevalence and distribution of disorders and their
correlates in particular populations of children
who vary in age, sex, socioeconomic status (SES),
ethnicity, or other characteristics (Costello &
Angold, 2000). The overall lifetime prevalence
rates for childhood problems are estimated to be
high and on the order of 14–22% of all children
(Rutter, 1989). Rutter, Tizard, and Whitmore
(1970), in the classic Isle of Wight Study, found
the overall rate of child psychiatric disorders to
be 6–8% in 9- to 11-year-old children. Richman,
Stevenson, and Graham (1975), in the London
Epidemiological Study, found moderate to severe
behavior problems for 7% of the population, with
an additional 15% of children having mild problems.
Boyle et al. (1987) and Offord et al. (1987),
in the Ontario Child Health Study, reported that
19% of boys and 17% of girls had one or more
disorders. Many other epidemiological studies
have reported similar rates of prevalence (e.g.,
Brandenburg, Friedman, & Silver, 1990; Costello,
Farmer, Angold, Burns, & Erkanli, 1997;
Earls, 1980; Hewitt et al., 1997; Lapouse &
Monk, 1958; MacFarlane, Allen, & Honzik, 1954;
Shaffer et al., 1996; Verhulst & Koot, 1992;
Werner, Bierman, & French, 1971). Perhaps the
most consistent general conclusions to be drawn
from these studies are that prevalence rates for
childhood problems are generally high, but that
rates vary with the nature of the disorder; the age,
sex, SES, and ethnicity of the child; the criteria
used to define the problem; the method used
to gather information (e.g., interview, questionnaire);
the informant (e.g., child, parent, teacher);
12 I. INTRODUCTION
sampling considerations; and a host of other
factors.
Age Differences
Bird, Gould, Yager, Staghezza, and Camino
(1989) reported no significant age differences for
children aged 4 to 16 years in the total number
of Diagnostic and Statistical Manual of Mental
Disorders, third edition (DSM-III) disorders diagnosed
at each age. Some studies of nonclinical
samples of children have found a general decline
in overall problems with age (e.g., Achenbach &
Edelbrock, 1981), whereas similar studies of clinical
samples have found an opposite trend (e.g.,
Achenbach, Howell, Quay, & Conners, 1991).
Some studies have reported interactions among
number of problems, age, sex of child, problem
type, clinical status, and source of information
(e.g., Simonoff et al., 1997). For example, Achenbach
et al. (1991) found that externalizing problems
showed a decline with age relative to internalizing
problems, but only for those children
who had been referred for treatment; Offord,
Boyle, and Racine (1989) found complex interactions
between age and sex of the child, with the
results also depending on whether the informant
was a child, parent, or teacher.
These and other findings raise numerous questions
concerning age differences in children’s
problem behaviors. Answers to even a seemingly
simple question such as “Do problem behaviors
decrease (or increase) with age?” are complicated
by (1) a lack of uniform measures of behavior that
can be used across a wide range of ages; (2) qualitative
changes in the expression of behavior with
development; (3) the interactions between age
and sex of the child; (4) the use of different informants;
(5) the specific problem behavior(s) of
interest; (6) the clinical status of the children
being assessed; and (7) the use of different diagnostic
criteria for children of different ages.
Notwithstanding these difficulties, both longitudinal
(MacFarlane et al., 1954) and crosssectional
(Achenbach & Edelbrock, 1981; Achenbach
et al., 1991) general population surveys
are informative in depicting changes in the proportions
of specific parent-, teacher-, or childreported
problem behaviors with age (e.g.,
“hyperactive,” “argues,” “cries”), as well as the
manner in which the age changes vary as a function
of problem type, sex, and clinical status of the
child. However, it should be emphasized that
general age trends are based on group statistics,
which may obscure the nonlinear and nonnormative
changes that often occur for individual
children. In addition, general surveys do not provide
information concerning the processes underlying
age changes. Studies of change in individual
children over time and of the context in which this
change occurs are needed if such processes are
to be understood (e.g., Bergman & Magnusson,
1997; Francis, Fletcher, Stuebing, Davidson, &
Thompson, 1991).
Socioeconomic Status
Although most children with mental health problems
are from the middle class, mental health
problems are overrepresented among the very
poor. It is estimated that 20% or more of children
in North America are poor, and that as many
|as 20% of children growing up in inner-city
poverty are impaired to some degree in their
social, behavioral, and academic functioning
(Duncan et al., 1994; Institute of Medicine, 1989;
Schteingart, Molnar, Klein, Lowe, & Hartmann,
1995).
Lower-SES children have been reported to
display more psychopathology and other problems
than upper-SES children (e.g., Keenan,
Shaw, Walsh, Delliquadri, & Giovannelli, 1997;
Samaan, 2000). However, although the reported
relationships between SES and child psychopathology
are statistically significant, the effects
are small and should be interpreted cautiously
(Achenbach et al., 1991). More importantly,
global estimates of SES often tell us little about
the associated processes through which SES
exerts its influence on a child. Knowledge of such
processes is needed to inform our understanding
of disorders. For example, the effects of SES on
aggression can be explained mostly by stressful
life events and by beliefs that are accepting
of aggression (Guerra, Tolan, Huesmann, Van
Acker, & Eron, 1995).
The impact of socioeconomic disadvantage on
children derives from the fact that SES is a composite
variable that includes many potential
sources of negative influence (Bradley, Corwyn,
McAdoo, & Garc?a Coll, 2001). In addition to low
income, low SES is often characterized by low
maternal education, a low level of employment,
single-parent status, parental psychopathology,
limited resources, and negative life events (e.g.,
poor nutrition, exposure to violence). Since overall
indices of SES may include one or more of
these variables in any given study, the relation1.
A Developmental–Systems Perspective 13
ship that is reported between SES and child
psychopathology may vary as a function of the
particular index used, as well as ethnic factors
(McLeod & Nonnemaker, 2000). In short,
SES may serve as a proxy or indicator of other
more active factors that influence risk for child
psychopathology.
Some research findings in child psychopathology
are confounded by a failure to control for
SES. For example, although physically abused
children show higher levels of externalizing problems
than nonabused children (Mash, Johnston,
& Kovitz, 1983), it is not clear that physical abuse
and externalizing problems are associated when
the effects of SES are controlled for (Cummings,
Hennessy, Rabideau, & Cicchetti, 1994; Wolfe &
Mosk, 1983). The relationships among SES,
maltreatment, and behavior disorders are further
complicated by other findings that the effects
of physical abuse on internalizing disorders
may be independent of SES, whereas the effects
of abuse on externalizing disorders may be dependent
on SES-related conditions (Okun,
Parker, & Levendosky, 1994).
Sex Differences
Although sex differences in the expression of psychopathology
have been formally recognized
since the time that Freud presented his views at
the beginning of the 20th century, psychopathology
in girls has received far less research attention
than psychopathology in boys (Bell-Dolan,
Foster, & Mash, in press; Eme, 1979). In the past,
many studies have either excluded girls from their
samples entirely or have examined all children
together without considering findings for girls
separately. For example, until recently there were
relatively few studies on disruptive behavior
disorders in girls (e.g., Moffitt, Caspi, Rutter, &
Silva, 2001; Silverthorn & Frick, 1999; Zoccolillo,
1993). This omission was related to the perception
that such disorders are much more common
in boys than in girls; to sampling biases in which
boys, who are more severely disruptive, are also
more likely to be referred and studied; and to the
use of inclusionary diagnostic criteria that were
derived and validated largely from studies with
boys (Spitzer, Davies, & Barkley, 1990).
Research has confirmed that there are important
differences in the prevalence, expression,
accompanying disorders, underlying processes,
outcomes, and developmental course of psychopathology
in boys versus girls (Eme, 1979, 1992;
Hops, 1995; Keenan & Shaw, 1997; Willcutt &
Pennington, 2000a; Zahn-Waxler, 1993). ADHD,
autism, childhood disruptive behavior disorders,
and learning and communication disorders are all
more common in boys than girls, whereas the
opposite is true for most anxiety disorders, adolescent
depression, and eating disorders (Hartung
& Widiger, 1998). Although these sex differences
are well established, their meaning is not well
understood. For example, it is difficult to determine
whether observed sex differences are a
function of referral or reporting biases, the way
in which disorders are currently defined, differences
in the expression of the disorder (e.g., direct
vs. indirect aggressive behavior), sex differences
in the genetic penetrance of disorders, or
sex differences in biological characteristics and
environmental susceptibilities. All are possible,
and there is a need for research into the processes
underlying observed differences. Clearly the
mechanisms and causes of sex differences may
vary for different disorders (e.g., ADHD vs. depression),
or for the same disorder at different
ages (e.g., child vs. adolescent obsessive–compulsive
disorder or early- vs. late-onset conduct
disorder).
Early research into sex differences focused
mainly on descriptive comparisons of the frequencies
of different problems for boys versus
girls at different ages. In general, differences in
problem behaviors between the sexes are small
in children of preschool age or younger (e.g.,
Briggs-Gowan, Carter, Skuban, & Horwitz, 2001;
Gadow, Sprafkin, & Nolan, 2001), but become
increasingly common with age. For example,
Weisz and Suwanlert (1989) studied children in
the United States and Thailand, and found that
boys were rated higher than girls on every problem
for which there was a significant sex difference—
including total problems, undercontrolled
problems, overcontrolled problems, and culturespecific
problems. Across cultures, boys have
been found to display more fighting, impulsivity,
and other uncontrolled behaviors than girls
(Olweus, 1979).
It has been found that boys show greater difficulties
than girls during early or middle childhood,
particularly with respect to ADHD and
disruptive behavior disorders (MacFarlane et al.,
1954). Girls’ problems may increase during adolescence,
with higher prevalence rates for depression
and dysphoric mood from midadolescence
through adulthood. For example, conduct disorder
and hyperactivity have been found to be
14 I. INTRODUCTION
more frequent in 12- to 16-year-old boys than
girls, whereas emotional problems have been
found to be more frequent for girls than boys in
this age group (Boyle et al., 1987; Offord et al.,
1987). In addition, early signs of aggression have
been found to predict later antisocial behavior for
boys but not for girls (Tremblay et al., 1992).
However, not all studies have reported significant
sex differences in overall rates of problem
behavior (e.g., Achenbach & Edelbrock, 1981;
Velez, Johnson, & Cohen, 1989), and even when
significant overall sex differences have been
found, they tend to be small and to account for
only a small proportion of the variance. It has also
been found that although there is a much larger
predominance of externalizing problems in boys
and of internalizing problems in adolescent girls
in samples of children who are referred for treatment,
sex differences in externalizing versus internalizing
problems are minimal in nonreferred
samples of children (Achenbach et al., 1991).
Comparisons of the behavioral and emotional
problems in boys versus girls over time can provide
useful information about sex-related characteristics.
However, taken in isolation, such global
comparisons do not address possible qualitative
differences in (1) expressions of psychopathology
in boys versus girls; (2) the processes underlying
these expressions; (3) the long-term consequences
of certain behaviors for boys versus girls;
and/or (4) the impact of certain environmental
events on boys versus girls (Zahn-Waxler, 1993).
As noted by Hops (1995), it seems likely that “the
pathways from childhood to adolescence and
adult pathology are age and gender specific and
that these differences may be the result of different
social contexts that nurture the development
of health or pathology for female and male individuals”
(p. 428). In addition to differential socialization
practices, there are likely to be differences
in the expression and outcome of psychopathology
in boys versus girls as a function of biologically
based differences. For example, in a study
of the psychophysiology of disruptive behavior in
boys versus girls, Zahn-Waxler, Cole, Welsh, and
Fox (1995) found that disruptive girls showed
high electrodermal responding relative to disruptive
boys and were also highly activated by a sadness
mood induction. These investigators suggested
that girls’ disruptive behavior may be more
closely connected than boys’ disruptive behavior
to experiences of anxiety. Other research has
found that increases in depression in females
during adolescence are related mostly to accompanying
changes in levels of estrogen and androgen
(Angold, Costello, Erkanli, & Worthman,
1999). It is also possible that for some disorders
(e.g., ADHD), girls may require a higher genetic
loading for the disorders than boys before the
disorders are likely to express themselves (Rhee,
Waldman, Hay, & Levy, 1999).
There may also be differences in the processes
underlying the expression of psychopathology
and distress in boys versus girls. For example,
findings suggest that the an adolescent’s emergent
sexuality may create special difficulties with
the parent of the opposite sex, and that distress
in adolescent males may be particularly disruptive
for mothers and daughters (Ge et al., 1995).
Others studies have found that daughters of depressed
mothers may be at greater risk than sons
for the development of internalizing disorders
(Gelfand & Teti, 1990) and that sons of fathers
showing avoidant patterns of adjustment to marital
distress may be particularly susceptible to internalizing
disorders (Katz & Gottman, 1993).
Finally, depression in adolescent females has
been found to be strongly associated with maternal
depression, whereas a lack of supportive early
care appears to be more strongly associated with
depression in adolescent males (Duggal, Carlson,
Sroufe, & Egeland, 2001).
It has also been found that the types of childrearing
environments predicting resilience to
adversity may differ for boys and girls. Resilience
in boys is associated with households in which
there is a male model (e.g., father, grandfather,
older sibling), structure, rules, and some encouragement
of emotional expressiveness. In contrast,
resilient girls come from households that combine
risk taking and independence with support
from a female caregiver (e.g., mother, grandmother,
older sister) (Werner, 1995).
Zahn-Waxler et al. (1995) refer to the “gender
paradox of comorbidities,” which is that although
the prevalence of disruptive behavior is lower in
females than in males, the risk of comorbid conditions
such as anxiety is higher in female samples.
In explaining this paradox, Zahn-Waxler et al.
(1995) suggest that girls’ heightened level of interpersonal
sensitivity, caring, and empathy may be
a protective factor in insulating them from developing
antisocial behavior. At the same time, girls’
overreceptivity to the plight of others, and their
reluctance to assert their own needs in situations
involving conflict and distress, may elevate their
risk for the development of internalizing problems.
However, the relations between gender and
1. A Developmental–Systems Perspective 15
comorbidity are likely to vary with the disorders
under consideration, the age of the child, the
source of referral, and other factors. For example,
in contrast to Zahn-Waxler et al. (1995), Biederman
et al. (2002) found that girls with ADHD had
a significantly lower rate of comorbid major depression
than did boys with ADHD.
Although findings relating to sex differences
and child psychopathology are complex, inconsistent,
and frequently difficult to interpret, the
cumulative findings from research strongly indicate
that the effects of gender are critical to
understanding the expression and course of most
forms of childhood disorder (Bell-Dolan et al., in
press; Kavanagh & Hops, 1994). It is particularly
important to understand the processes and
mechanisms underlying these gender effects, and
to recognize that biological influences and differential
socialization practices are likely to interact
throughout development in accounting for any
differences between the sexes that are found.
Rural versus Urban Differences
Although there is a general belief that rates of
child behavior disorder are higher in urban than
in rural areas, research findings in support of this
view are weak and/or inconsistent. Findings from
the Isle of Wight, Inner London Borough, and
Ontario Child Health Studies reveal prevalence
rates of problem behavior that were higher for
urban than rural children (Offord et al., 1987;
Rutter, 1981). On the other hand, in a crosscultural
investigation, Weisz and Suwanlert
(1991) found few differences in parent or teacher
ratings of child problems as a function of rural
versus urban status in either of the cultures that
were studied (United States and Thailand). In a
detailed analysis that controlled for the effects of
SES and ethnicity and also looked at gradations
of urbanization, Achenbach et al. (1991) found
few differences in children’s behavior problems
or competencies as a function of rural versus
urban status, although there was a significant but
very small effect indicating higher delinquency
scores for children in urban environments. These
investigators concluded that earlier findings of
higher rates of problem behavior in urban than
in rural areas “may have reflected the tendency
to combine areas of intermediate urbanization with
large urban areas for comparison with rural areas
as well as a possible lack of control for demographic
differences” (p. 86). Even in studies in which rural
versus urban differences have been found, for the
most part these differences were associated with
economic and cultural differences between sites,
and not with urbanization per se (Zahner, Jacobs,
Freeman, & Trainor, 1993).
Ethnicity and Culture
Ethnicity
Numerous terms have been used to describe ethnic
influences. These include “ethnicity,” “race,”
“ethnic identity,” “ethnic orientation,” “acculturation,”
“bicultural orientation,” and “culture.” As
pointed out by Foster and Martinez (1995), there
is a need to recognize the diversity of terminology
that has been used in describing ethnicity,
and the fact that these terms refer to related but
different things. Despite the growing ethnic
diversity of the North American population,
ethnic representation in research studies and
the study of ethnicity-related issues more generally
have received relatively little attention in
studies of child psychopathology (Garc?a Coll,
Akerman, & Cicchetti, 2000; U.S. Public Health
Service, 2001c). In lamenting this state of affairs,
Foster and Martinez (1995) state: “The underrepresentation
of children from diverse backgrounds
is accompanied by a dearth of empirical
literature on the origins, correlates, and treatment
of child psychopathology in different ethnic
groups within the United States. Instead, investigators
have based theories of child behavior,
both normal and deviant, on data drawn largely
from European-American culture” (p. 214).
Research into child psychopathology has generally
been insensitive to possible differences in
prevalence, age of onset, developmental course,
and risk factors related to ethnicity (Kazdin &
Kagan, 1994), and to the considerable heterogeneity
that exists within specific ethnic groups
(Murry, Bynum, Brody, Willert, & Stephens,
2001; Murry, Smith, & Hill, 2001). In addition,
few studies have compared ethnic groups while
controlling for other important variables, such as
SES, sex, age, and geographic region. In recent
comparisons that have controlled for these variables,
African American and Hispanic American
children are identified and referred at the same
rates as other children, but they are much less
likely to actually receive specialty mental health
services or psychotropic medications (Garc?a Coll
& Garrido, 2000). European American and Native
American children have been found to display
similar mental health problems with the
16 I. INTRODUCTION
exception of substance abuse, where rates are
higher for Native American youngsters (Costello,
Farmer, & Angold, 1999).
Some studies that have included a small number
of African American children in their samples
have reported somewhat higher rates of externalizing
problems for this group (Costello, 1989;
Velez et al., 1989). However, other studies with
much larger national samples that included European
American, African American, and Hispanic
American children have reported either no or
very small differences related to race or ethnicity
when SES, sex, age, and referral status were
controlled for (Achenbach & Edelbrock, 1981;
Achenbach et al., 1991; Lahey et al., 1995). So,
although externalizing problems have been reported
to be more common among African
American children, this finding is probably an
artifact related to SES. Externalizing disorder is
associated with both ethnicity and SES, and since
there is an overrepresentation of minority status
children in low-SES groups in North America,
caution must be exercised in interpreting the relationships
among SES, ethnicity, and aggression
(Guerra et al., 1995; Lahey et al., 1995).
Ethnicity has not been found to be strongly
associated with risk for eating disorders (Leon,
Fulkerson, Perry, & Early-Zald, 1995), although
differences between European Americans and
other groups have been reported for such subclinical
eating disturbances as dietary restraint,
ideal body shape, and body dissatisfaction (Wildes
& Emery, 2001). Differing patterns of substance
abuse as a function of ethnicity have also been
reported (Catalano et al., 1993). More research
is needed, but these and other findings suggest
that the effects of ethnicity are likely to vary with
the problem under consideration and its severity.
As is the case for SES and sex differences,
global comparisons of the prevalence of different
types of problems for different ethnic groups are
not likely to be very revealing. On the other hand,
studies into the processes affecting the form, associated
factors, and outcomes of different disorders
for various ethnic groups hold promise for
increasing our understanding of the relationship
between ethnicity and child psychopathology (e.g.,
Bird et al., 2001; Bradley, Corwyn, Burchinal,
McAdoo, & Garc?a Coll, 2001).
Culture
The values, beliefs, and practices that characterize
a particular ethnocultural group contribute to
the development and expression of childhood
distress and dysfunction, which in turn are organized
into categories through cultural processes
that further influence their development and
expression (Harkness & Super, 2000; Wong &
Ollendick, 2001). Through shared views about
causality and intervention, culture also structures
the way in which people and institutions react to
a child’s problems. Since the meaning of children’s
social behavior is influenced by cultural
beliefs and values, it is not surprising that the
form, frequency, and predictive significance of
different forms of child psychopathology vary
across cultures, or that cultural attitudes influence
diagnostic and referral practices (Lambert
& Weisz, 1992). For example, shyness and oversensitivity
in children have been found to be associated
with peer rejection and social maladjustment
in Western cultures, but with leadership,
school competence, and academic achievement
in Chinese children in Shanghai (Chen, Rubin,
& Li, 1995). Similarly, Lambert and Weisz (1989)
found that overcontrolled problems were reported
significantly more often for Jamaican than
for American youngsters—a finding consistent
with Afro-British Jamaican cultural attitudes and
practices that discourage child aggression and
other undercontrolled behavior, and that foster
inhibition and other overcontrolled behavior.
Weisz and Sigman (1993), using parent reports
of behavioral and emotional problems in 11- to
15-year-old children from Kenya, Thailand, and
the United States, found that Kenyan children
were rated particularly high on overcontrolled
problems (e.g., fears, feelings of guilt, somatic
concerns), due primarily to numerous reports of
somatic problems. In this mixed-race sample,
whites were rated particularly high on undercontrolled
problems (e.g., “arguing,” “disobedient
at home,” “cruel to others”). Weisz and Suwanlert
(1987) compared 6- to 11-year-old children in the
Buddhist-oriented, emotionally controlled culture
of Thailand with American 6- to 11-yearolds.
Parent reports revealed Thai–U.S. differences
in 54 problem behaviors, most of which
were modest in magnitude. Thai children were
rated higher than American children on problems
involving overcontrolled behaviors such as anxiety
and depression, whereas American children
were rated higher than Thai children on undercontrolled
behaviors such as disobedience and
fighting.
Weisz and Suwanlert (1991) compared ratings
of behavior and emotional problems of 2- to
1. A Developmental–Systems Perspective 17
9-year-old children in Thailand and the United
States. Parents and teachers in Thailand rated
both overcontrolled and undercontrolled problems
as less serious, less worrisome, less likely to
reflect personality traits, and more likely to improve
with time. These findings suggest that there
may be cultural differences in the meanings ascribed
to problem behaviors across cultures.
Findings from these and other studies suggest
that the expression of, and tolerance for, many
child behavioral and emotional disturbances are
related to social and cultural values. The processes
that mediate this relationship are in need
of further investigation. In this regard, it is important
that research on child psychopathology
not be generalized from one culture to another,
unless there is support for doing so. There is some
support for the notion that some processes—for
example, those involved in emotion regulation
and its relation to social competence—may be
similar across diverse cultures (Eisenberg, Pidada,
& Liew, 2001). The rates of expression of some
disorders, particularly those with a strong neurobiological
basis (e.g., ADHD, autistic disorder),
may be less susceptible to cultural influences than
others. However, even so, social and cultural beliefs
and values are likely to influence the meaning
given to these behaviors, the ways in which
they are responded to, their forms of expression,
and their outcomes.
An important distinction to be made with respect
to cross-cultural comparisons is whether or
not there are real differences in the rates of the
disorder, or differences in the criteria used to
make judgments about these problems. For example,
Weisz and Suwanlert (1989) compared
the teacher-reported behavioral/emotional problems
of Thai and U.S. children (ages 6–11 years).
It was found that Thai teachers were confronted
with students who were more prone to behavioral
and emotional problems at school than were teachers
in the United States, but that they applied different
judgments to the behaviors they observed.
Cultural factors are known to influence not
only informal labeling processes but formal diagnostic
practices as well. For example, reported
prevalence rates of ADHD in Britain are much
lower than in the United States, because of differences
in the way in which diagnostic criteria
for ADHD are applied in the two countries. Such
differences in diagnostic practices may lead to
spurious differences in reported prevalence rates
for different forms of child psychopathology
across cultures.
Cross-cultural research on child psychopathology
would suggest that the expression and experience
of mental disorders in children are not
universal (Fisman & Fisman, 1999). Patterns of
onset and duration of illness and the nature and
relationship among specific symptoms vary from
culture to culture, and across ethnic groups
within cultures (Hoagwood & Jensen, 1997).
However, few studies have compared the attitudes,
behaviors, and biological and psychological
processes of children with mental disorders
across different cultures. Such information is
needed to understand how varying social experiences
and contexts influence the expression,
course, and outcome of different disorders across
cultures. For example, greater social connectedness
and support in more traditional cultures, and
greater access to resources and opportunities in
industrialized societies, are examples of mechanisms
that may alter outcomes across cultures.
Sensitivity to the role of cultural influences in
child psychopathology has increased (Evans &
Lee, 1998; Lopez & Guarnaccia, 2000), and is
likely to continue to do so as globalization and
rapid cultural change become increasingly more
common (Garc?a Coll et al., 2000).
BASIC ISSUES IN CHILD
PSYCHOPATHOLOGY
Several recurrent and overlapping issues have
characterized the study of psychopathology in
children (Rutter & Garmezy, 1983; Rutter &
Sroufe, 2000). A number of these are highlighted
in this section, including (1) difficulties in conceptualizing
psychopathology and normality; (2) the
need to consider healthy functioning and adjustment;
(3) questions concerning developmental
continuities and discontinuities; (4) the concept
of developmental pathways; (5) the notions of risk
and resilience; (6) the identification of protective
and vulnerability factors; and (7) the role of contextual
influences.
Psychopathology versus Normality
Conceptualizing child psychopathology and attempting
to establish boundaries between what
constitutes abnormal and normal functioning are
arbitrary processes at best (Achenbach, 1997).
Traditional approaches to mental disorders in
children have emphasized concepts such as symptoms,
diagnosis, illness, and treatment; by doing
18 I. INTRODUCTION
so, they have strongly influenced the way we think
about child psychopathology and related questions
(Richters & Cicchetti, 1993). Childhood
disorders have most commonly been conceptualized
in terms of deviancies involving breakdowns
in adaptive functioning, statistical deviation,
unexpected distress or disability, and/or
biological impairment.
Wakefield (1992, 1997, 1999a) has proposed
an overarching concept of mental disorder as
“harmful dysfunction.” This concept encompasses
a child’s physical and mental functioning,
and includes both value- and science-based criteria.
In the context of child psychopathology,
a child’s condition is viewed as a disorder only if
(1) it causes harm or deprivation of benefit to the
child, as judged by social norms; and (2) it results
from the failure of some internal mechanism to
perform its natural function (e.g., “an effect that
is part of the evolutionary explanation of the
existence and structure of the mechanism”;
Wakefield, 1992, p. 384). This view of mental disorder
focuses attention on internally evolved
mechanisms—for example, executive functions in
the context of self-regulation (Barkley, 2001).
Nevertheless, as pointed out by Richters and
Cicchetti (1993), this view only identifies the
decisions that need to be made in defining mental
disorders; it does not specify how such decisions
are to be made.
As is the case for most definitions of mental
disorder that have been proposed, questions related
to defining the boundaries between normal
and abnormal, understanding the differences
between normal variability and dysfunction, defining
what constitute “harmful conditions,” linking
dysfunctions causally with these conditions,
and circumscribing the domain of “natural” or of
other proposed mechanisms are matters of considerable
controversy (Lilienfeld & Marino, 1995;
Richters & Cicchetti, 1993).4 Categories of mental
disorder stem from human-made linguistic
distinctions and abstractions, and boundaries
between what constitutes normal and abnormal
conditions, or between different abnormal conditions,
are not easily drawn. Although it may
sometimes appear that efforts to categorize mental
disorders are carving “nature at its joints,”
whether or not such “joints” actually exist is open
to debate (e.g., Cantor, Smith, French, &
Mezzich, 1980; Lilienfeld & Marino, 1995). However,
clear joints do not necessarily need to exist
for categorical distinctions to have utility. For
instance, there is no joint at which one can carve
day from night; yet distinguishing the two has
proven incredibly useful to humans in going
about their social discourse and engagements.
Likewise, although the threshold for determining
disorder from merely high levels of symptoms
may be fuzzy, it could be stipulated as being at
that point along a dimension where impairment
in a major, culturally universal life activity befalls
the majority of people at or exceeding that point.
Thus, despite the lack of clear boundaries between
what is normal and abnormal, categorical
distinctions are still useful.
Healthy Functioning
The study of psychopathology in children requires
concomitant attention to adaptive developmental
processes for several reasons. First,
judgments of deviancy require knowledge of normative
developmental functioning, both with respect
to a child’s performance relative to sameage
peers and with respect to the child’s own
baseline of development. Second, maladaptation
and adaptation often represent two sides of the
same coin, in that dysfunction in a particular
domain of development (e.g., the occurrence of
inappropriate behaviors) is usually accompanied
by a failure to meet developmental tasks and
expectations in the same domain (e.g., the nonoccurrence
of appropriate behaviors). It is important
to point out, however, that adaptation should
not be equated with the mere absence of psychopathology.
Kendall and his colleagues (Kendall,
Marrs-Garcia, Nath, & Sheldrick, 1999; Kendall
& Sheldrick, 2000), for instance, contend that
it is important to use normative comparisons to
evaluate treatment outcome; they suggest that
improvement involves falling within a certain
range of healthy functioning, in addition to the
amelioration of one’s symptom presentation.
Moreover, adaptation involves the presence and
development of psychological, physical, interpersonal,
and intellectual resources (see Fredrickson,
2001). Third, in addition to the specific problems
that lead to referral and diagnosis, disturbed children
are likely to show impairments in other areas
of adaptive functioning. For example, in addition
to their core symptoms of impulsivity and inattention,
children with ADHD also show lower-thanaverage
levels of functioning in their socialization,
communication, and activities of daily living (e.g.,
Stein, Szumowski, Blondis, & Roizen, 1995).
Fourth, most children with specific disorders are
known to cope effectively in some areas of their
1. A Developmental–Systems Perspective 19
lives. Understanding a child’s strengths informs
our knowledge of the child’s disorder and provides
a basis for the development of effective treatment
strategies. Fifth, children move between pathological
and nonpathological forms of functioning
over the course of their development. Individual
children may have their “ups and downs” in problem
type and frequency over time. Sixth, many
child behaviors that are not classifiable as deviant
at a particular point in time may nevertheless represent
less extreme expressions or compensations
of an already existing disorder or early expressions
of a later progression to deviant extremes as development
continues (Adelman, 1995). Finally, no
theory of a childhood disorder is complete if it
cannot be linked with a theory of how the underlying
normal abilities develop and what factors go
awry to produce the disordered state. Therefore,
understanding child psychopathology requires
that we also attend to these less extreme forms of
difficulty and develop more complete models of
the normal developmental processes underlying
the psychopathology.
For these and other reasons to be discussed,
the study of child psychopathology requires an
understanding of both abnormal and healthy
functioning. As noted by Cicchetti and Richters
(1993), “it is only through the joint consideration
of adaptive and maladaptive processes within the
individual that it becomes possible to speak in
meaningful terms about the existence, nature,
and boundaries of the underlying psychopathology”
(p. 335). To date, far greater attention has
been devoted to the description and classification
of psychopathology in children than to healthy
child functioning; to nonpathological psychosocial
problems related to emotional upset, misbehavior,
and learning; or to factors that promote
the successful resolution of developmental tasks
(Adelman, 1995; Sonuga-Barke, 1998). In light
of this imbalance, there is a need for studies of
normal developmental processes (Lewis, 2000),
for investigations of normative and representative
community samples of children (Ialongo, Kellam,
& Poduska, 2000; Kazdin, 1989), and for studies
of “resilient” children who show normal development
in the face of adversity (Masten, 2001).
Developmental Continuities
and Discontinuities
A central issue for theory and research in child
psychopathology concerns the continuity of disorders
identified from one time to another and the
relationship between child, adolescent, and
adult disorders (Caspi, 2000; Garber, 1984;
Kazdin & Johnson, 1994; Rutter & Rutter, 1993;
Sroufe & Jacobvitz, 1989). Over the past two
decades, research into early attachment has
stimulated general interest in the roles of relational
processes and internalized representational
systems as the bases for understanding continuities
and discontinuities in psychopathology over
time and across generations (Cassidy & Shaver,
1999; Lyons-Ruth, 1995; Sroufe, Duggal, Weinfeld,
& Carlson, 2000).
Some childhood disorders, such as mental retardation
and autistic disorder, are chronic conditions
that will persist throughout childhood and
into adulthood. Other disorders, such as functional
enuresis and encopresis, occur during
childhood and only rarely manifest themselves in
adults (Walker et al., 1989). And other disorders
(e.g., mood disorders, schizophrenia, generalized
anxiety disorder) are expressed, albeit in modified
forms, in both childhood and adulthood and
exhibit varying degrees of continuity over time.
Evidence in support of the continuity between
child and adult disorders is equivocal and depends
on a number of methodological factors
related to research design, assessment instruments,
the nature of the study sample, and the
type and severity of the disorder (Garber, 1984).
In general, the literature suggests that child psychopathology
is continuous with adult disorders
for some, but not all, problems. As we discuss
below, there is evidence that appears to favor the
stability of externalizing problems over internalizing
problems. However, previous findings may
reflect the severity and pervasiveness of the disorders
assessed, referral biases, and the fact that
longitudinal investigations of children with internalizing
and other disorders are just beginning to
emerge. For example, one study found that firstgrade
anxious symptoms predicted levels of anxious
symptoms and adaptive functioning in fifth
grade (Ialongo, Edelsohn, Werthamer-Larsson,
Crockett, & Kellam, 1995). In another report,
early-onset bulimia nervosa was associated with
a 9-fold increase in risk for late-adolescent bulimia
nervosa and a 20-fold increase in risk for
adult bulimia nervosa (Kotler et al., 2001).
The possible mechanisms underlying the relationships
between early maladaptation and later
disordered behavior are numerous and can operate
in both direct and indirect ways (Garber, 1984;
Rutter, 1994a; Sroufe & Rutter, 1984). Some examples
of direct relationships between early and
20 I. INTRODUCTION
later difficulties include (1) the development of
a disorder during infancy or childhood, which
then persists over time; (2) experiences that alter
the infant’s or child’s physical status (e.g., neural
plasticity), which in turn influences later functioning
(Courchesne, Chisum, & Townsend, 1994;
Johnson, 1999; Nelson, 2000); and (3) the acquisition
of early patterns of responding (e.g., compulsive
compliance, dissociation) that may be
adaptive in light of the child’s current developmental
level and circumstances, but may result
in later psychopathology when circumstances
change and new developmental challenges arise.
Some examples of indirect associations between
child and adult psychopathology may involve early
predispositions that eventually interact with environmental
experiences (e.g., stressors), the combination
of which leads to dysfunction. For example,
Egeland and Heister (1995) found that
the impact of day care on disadvantaged high-risk
children at 42 months of age was related to the
children’s attachment quality at 12 months of age,
with securely attached children more likely to be
negatively affected by early out-of-home care.
Other examples of indirect links between child
and adult disturbance include (1) experiences
(e.g., peer rejection) that contribute to an altered
sense of self-esteem (DuBois & Tevendale, 1999),
or that create a negative cognitive set, which then
leads to later difficulties; and (2) experiences providing
various opportunities or obstacles that then
lead to the selection of particular environmental
conditions, and by doing so guide a child’s course
of development (Rutter, 1987; Sroufe & Rutter,
1984).
Research efforts have focused not only on the
continuities and discontinuities in childhood disorders,
but also on the identification of factors
that predict them. One factor that has been studied
in the context of conduct disorder is age of
onset, with early onset usually viewed as the occurrence
of conduct disorder symptoms prior to
age 12 years (Loeber & Dishion, 1983; O’Donnell,
Hawkins, & Abbott, 1995). It has been found that
early onset of symptoms is associated with higher
rates and more serious antisocial acts over a
longer period of time for both boys and girls
(Lavigne et al., 2001). However, psychosocial
variables that are present prior to and following
onset may influence the seriousness and chronicity
more than age of onset per se does (Tolan &
Thomas, 1995). A question that needs to be addressed
is this: Does early age of onset operate
in a causal fashion for later problems, and if so,
how? Another issue is whether the causal processes
that are associated with an early onset of
a disorder (e.g., depression) are different from
those that serve to maintain the disorder. Even
then, the specification of an age of onset need not
be made so precisely that it creates a false distinction
that only valid cases meet that precise threshold,
as may have happened with ADHD (see
Barkley, chapter 2, this volume). Such efforts to
impose precision where none exists may have
backfired in hampering studies of teens and
adults having the same disorder who cannot adequately
recall such a precise onset, and in presuming
that cases having qualitatively identical
symptoms and impairments but later onsets are
invalid instances of a disorder.
Although research supports the notion of continuity
of disorders, it does not support the continuity
of identical symptoms over time (i.e.,
“homotypic correspondence”). Continuity over
time for patterns of behavior rather than for specific
symptoms is the norm. For example, although
externalizing disorders in boys are stable
over time, the ways in which these behavioral patterns
are expressed are likely to change dramatically
over the course of development (Olweus,
1979). Even with wide fluctuations in the expression
of behavior over time, “children may show
consistency in their general adaptive or maladaptive
pattern of organizing their experiences and
interacting with the environment” (Garber, 1984,
p. 34). Several research findings can be used to
illustrate this notion of consistent “patterns of
organization.” For example, early heightened
levels of behavioral inhibition may affect later adjustment
by influencing the way in which a child
adapts to new and unfamiliar situations and the
ensuing person–environment interactions over
time (Kagan, 1994a). Another example of a consistent
pattern of organization involves early attachment
quality and the development of internal
working models that children carry with them
into their later relationships (Bowlby, 1988;
Goldberg, 1991). Internal working models of self
and relationships may remain relatively stable
over time, at the same time that the behavioral
expressions of these internal models change with
development. From a neuroscientific perspective,
Pennington and Ozonoff (1991) argue that
certain genes and neural systems also play a significant
predisposing role in influencing the continuity
of psychopathology, and that the “discontinuities
at one level of analyses—that of
observable behavior—may mask continuities at
1. A Developmental–Systems Perspective 21
deeper levels of analysis; those concerned with
the mechanisms underlying observable behavior”
(p. 117).
Given that developmental continuity is reflected
in general patterns of organization over
time rather than in isolated behaviors or symptoms,
the relationships between early adaptation
and later psychopathology are not likely to be
direct or uncomplicated. The connections between
psychopathology in children and adults are
marked by both continuities and discontinuities.
The degree of continuity–discontinuity will vary
as a function of changing environmental circumstances
and transactions between a child and
the environment that affect the child’s developmental
trajectory.
Developmental Pathways
The concept of “developmental pathways” is crucial
for understanding continuities and discontinuities
in psychopathology. Such pathways are
not directly observable, but function as metaphors
that are inferred from repeated assessments
of individual children over time (Loeber,
1991). A pathway, according to Loeber (1991),
“defines the sequence and timing of behavioral
continuities and transformations and, ideally,
summarizes the probabilistic relationships between
successive behaviors” (p. 98). In attempting
to identify developmental pathways as either
“deviant” or “normal,” it is important to recognize
that (1) different pathways may lead to similar
expressions of psychopathology (i.e., “equifinality”);
and (2) similar initial pathways may
result in different forms of dysfunction (i.e.,
“multifinality”), depending on the organization of
the larger system in which they occur (Cicchetti
& Rogosch, 1996; Lewis, 2000; Loeber, 1991).
Research findings related to child maltreatment
provide an example of a possible developmental
pathway. It has been found that physically
abused children are more likely to develop insecure
attachments, view interpersonal relationships
as being coercive and threatening, become
vigilant and selectively attend to hostile cues, instantly
classify others as threatening or nonthreatening,
and acquire aggressive behavioral strategies
for solving interpersonal problems (see
Cicchetti & Manly, 2001). These children bring
representational models to peer relationships that
are negative, conflictual, and unpredictable. They
process social information in a biased and deviant
manner, and develop problems with peer
relationships that involve social withdrawal, unpopularity,
and overt social rejection by peers
(Dodge, Pettit, & Bates, 1994). In another example
of a developmental pathway, the diagnosis
of conduct disorder typically precedes the initiation
of use of various substances, and this use
in turn precedes the diagnosis of alcohol dependence
in adolescents (Kuperman et al., 2001).
The systematic delineation of developmental
pathways not only offers several advantages for
the study of the etiology and outcomes of childhood
disorders, but may also suggest strategies
for intervention. Loeber (1991, p. 99) describes
these advantages as “attempts to capture the
changing manifestations and variable phenotype
of a given disorder” over time. In this way, the
study of developmental pathways includes etiological
considerations, the assessment of comorbidities
as they accrue over time, and a sensitivity
to diverse outcomes (e.g., White, Bates, &
Buyske, 2001).
Risk and Resilience
Previous studies of child psychopathology focused
on elucidating the developmental pathways
for deviancy and maladjustment to the relative
exclusion of those for competency and adjustment
(but see Luthar, 1993; Rutter, 1985, 1987,
1994b; and Rutter & Rutter, 1993, for exceptions).
However, a significant number of children
who are at risk do not develop later problems.
There is a growing recognition of the need to
examine not only risk factors, but also those
conditions that protect vulnerable children from
dysfunction and lead to successful adaptations
despite adversity (Cicchetti & Garmezy, 1993).
“Resilience,” which refers to successful adaptations
in children who experience significant
adversity, has now received a good deal of attention
(Luthar, Cicchetti, & Becker, 2000). Early
patterns of adaptation influence later adjustment
in complex and reciprocal ways. Adverse conditions,
early struggles to adapt, and failure to meet
developmental tasks do not inevitably lead to a
fixed and unchanging abnormal path. Rather,
many different factors, including chance events
and encounters, can provide turning points
whereby success in a particular developmental
task (e.g., educational advances, peer relationships)
shifts a child’s course onto a more adaptive
trajectory. Conversely, there are numerous
events and circumstances and underlying dynamic
biological systems that may deflect the child’s
22 I. INTRODUCTION
developmental trajectory toward maladaptation
(e.g., a dysfunctional home environment, peer
rejection, difficulties in school, parental psychopathology,
intergenerational conflict, and even
late-onset genetic effects).
Although the term “resilience” has not been
clearly operationalized, it is generally used to describe
children who (1) manage to avoid negative
outcomes and/or to achieve positive outcomes despite
being at significant risk for the development
of psychopathology; (2) display sustained competence
under stress; or (3) show recovery from
trauma (Werner, 1995). Risk is usually defined in
terms of child characteristics that are known to be
associated with negative outcomes—for example,
difficult temperament (Ingram & Price, 2001;
Rothbart, Ahadi, & Evans, 2000)—and/or in terms
of a child’s exposure to extreme or disadvantaged
environmental conditions (e.g., poverty or abuse).
Individual children who are predisposed to develop
psychopathology and who show a susceptibility to
negative developmental outcomes under high-risk
conditions are referred to as “vulnerable.” Genetic
makeup and temperament are two factors that are
presumed to contribute to susceptibility for children
who are exposed to high-risk environments
(Rutter, 1985; Seifer, 2000).
Research on resilience has lacked a consistent
vocabulary, conceptual framework, and methodological
approach (Luthar et al., 2000; Rutter,
2000c; Zimmerman & Arunkumar, 1994). It is
particularly important to ensure that resilience is
not defined as a universal, categorical, or fixed
attribute of a child, but rather as a number of
different types of dynamic processes that operate
over time. Individual children may be resilient
in relation to some specific stressors but not
others, and resilience may vary over time and
across contexts (Freitas & Downey, 1998). As
noted by Zimmerman and Arunkumar (1994,
p. 4), “research on resiliency can only identify
those particular risk circumstances when environmental
conditions, individual factors, and
developmental tasks interact to help children
and adolescents avoid negative consequences.”
Fortunately, models of resilience have increasingly
begun to address the complex and dynamic
relationships between the child and his or her
environment, to incorporate the theoretical and
empirical contributions of developmental psychology,
and to acknowledge the multiple factors
related to normal and deviant behavior (Glantz
& Johnson, 1999; Walden & Smith, 1997; Tebes,
Kaufman, Adnopoz, & Racusin, 2001).
One problem in research on resilience has
been an absence of agreed-upon criteria for defining
positive developmental outcomes (see
Kaufman, Cook, Arny, Jones, & Pittinsky, 1994,
for a review of the ways in which positive outcomes
in studies of resilience have been operationalized).
For example, there is currently debate
as to whether the criteria for defining
resilience and adaptation should be based on
evidence from external criteria (e.g., academic
performance), internal criteria (e.g., subjective
well-being), or some combination of these (see
Masten, 2001). Variations across studies in the
source of information (e.g., parent or teacher);
the type of assessment method (e.g., interview,
questionnaire, observation); the adaptational criteria
used; and the number and timing of assessments
can easily influence the proportion of children
who are designated as resilient or not in any
particular investigation (Kaufman et al., 1994;
Masten, 2001). And there is also some confusion
about and circularity in how the term “resilience”
has been used, in that it has been used to refer to
both an outcome and to the cause of an outcome.
Several different models of resilience have also
been proposed, the most common ones being a
compensatory model, a challenge model (e.g.,
stress inoculation), and a protective-factors model
(Garmezy, Masten, & Tellegen, 1984).
Years of research suggest that resilience is not
indicative of any rare or special qualities of the
child per se (as implied by the term “the invulnerable
child”), but rather is the result of the interplay
of normal developmental processes such
as brain development, cognition, caregiver–child
relationships, regulation of emotion and behavior,
and the motivation for learning (Masten,
2001). Some researchers have argued that resilience
may be more ubiquitous than previously
thought, and that this phenomenon is part of the
“ordinary magic” and makeup of basic human
adaptation (Masten, 2001; Sheldon & King,
2001). It is when these adaptational systems are
impaired, usually through prolonged or repeated
adversity, that the risk for childhood psychopathology
increases.
Protective and Vulnerability Factors
Various protective and vulnerability factors have
been found to influence children’s reactions to
potential risk factors or stressors. These include
factors within the child, the family, and the community
(Osofsky & Thompson, 2000; Werner &
1. A Developmental–Systems Perspective 23
Smith, 1992). Common risk factors that have been
found to have adverse effects on a child encompass
both acute stressful situations and chronic
adversity; they include such events as chronic poverty,
serious caregiving deficits, parental psychopathology,
death of a parent, community disasters,
homelessness, reduced social support, decreased
financial resources, family breakup, parental marital/
couple conflict, and perinatal stress (Deater-
Deckard & Dunn, 1999; Rutter, 1999; Tebes et al.,
2001; Walden & Smith, 1997).
Protective factors within a child that have been
identified include an “easy” temperament (i.e., a
child who is energetic, affectionate, cuddly, goodnatured,
and/or easy to deal with), which makes
the child engaging to other people; early coping
strategies that combine autonomy with help seeking
when needed; high intelligence and scholastic
competence; effective communication and
problem-solving skills; positive self-esteem and
emotions; high self-efficacy; and the will to be or
do something (Fredrickson, 2001; Gilgun, 1999;
Werner, 1995). An example of a possible protective
factor within the child is seen in findings that
high vagal tone and vagal suppression—taken as
indices of a child’s ability to regulate emotion via
self-soothing, focused attention, and organized and
goal-directed behavior—can buffer children from
the increases in externalizing behaviors, internalizing
behaviors, and social problems often associated
with exposure to parental marital/couple hostility
and discord (Katz & Gottman, 1995) or
parental problem drinking (El-Sheikh, 2001).
At a family level, protective factors that have
been identified include the opportunity to establish
a close relationship with at least one person
who is attuned to the child’s needs, positive
parenting, availability of resources (e.g., child
care), a talent or hobby that is valued by adults
or peers, and family religious beliefs that provide
stability and meaning during times of hardship or
adversity (Werner & Smith, 1992). Protective
factors in the community include extrafamilial
relationships with caring neighbors, community
elders, or peers; an effective school environment,
with teachers who serve as positive role models
and sources of support; and opening of opportunities
at major life transitions (e.g., adult education,
voluntary military service, church or community
participation, a supportive friend or
marital/relationship partner).
In summary, early patterns of adaptation influence
later adjustment in complex and reciprocal
ways. Adverse conditions, early adaptational
struggles, and failure to meet developmental tasks
do not inevitably lead to a fixed and unmalleable
dysfunctional path. Rather, as noted earlier, many
different factors can act to alter a child’s developmental
course for the better. Conversely,
numerous events and circumstances may serve to
alter this course for the worse.
The interrelated issues of developmental continuities–
discontinuities, developmental pathways,
risk and resilience, and vulnerability and
protective factors are far from being resolved or
clearly understood. The multitude of interdependent
and reciprocal influences, mechanisms, and
processes involved in the etiology and course of
child psychopathology clearly suggest a need for
more complex theories (e.g., chaos theory, nonlinear
dynamic models) (Barton, 1994; Glantz &
Johnson, 1999; Gottman, Guralnick, Wilson,
Swanson, & Murray, 1997; Haynes & Blaine,
1995), research designs, and data-analytic strategies
(Kazdin & Kagan, 1994; Mash & Krahn, 2000;
Richters, 1997).
Contextual Influences
Messick (1983) cogently argues that any consideration
of child psychopathology must consider
and account for three sets of contextual
variables: (1) the child as context—the idea that
unique child characteristics, predispositions, and
traits influence the course of development; (2) the
child of context—the notion that the child comes
from a background of interrelated family, peer,
classroom, teacher, school, community, and cultural
influences; and (3) the child in context—the
understanding that the child is a dynamic and
rapidly changing entity, and that descriptions
taken at different points in time or in different
situations may yield very different information.
Research has increasingly come to recognize
the reciprocal transactions between the developing
child and the multiple social and environmental
contexts in which development occurs
(Cicchetti & Aber, 1998; Deater-Deckard, 2001).
Understanding context requires a consideration
of events that impinge directly on the child in a
particular situation at a particular point in time;
extrasituational events that affect the child indirectly
(e.g., a parent’s work-related stress); and
temporally remote events that continue to affect
the child through their representation in the
child’s current cognitive–affective data base.
Defining context has been, and continues to
be, a matter of some complexity. The context of
24 I. INTRODUCTION
maltreatment provides an illustration of difficulties
in definition. Maltreatment can be defined
in terms of its type, timing, frequency, severity,
and chronicity in the family (e.g., Manly, Kim,
Rogosch, & Cicchetti, 2001). Each of these
parameters and their interaction may contribute
to child outcomes, but in different ways. For example,
Manly, Cicchetti, and Barnett (1994) studied
different types of maltreatment and found
that outcomes generally did not differ for children
who were categorized as neglected versus
abused. However, a regression analysis indicated
that neglect accounted for more of the variance
in child problems than other types of abuse did.
In this study, sexually abused children were also
found to be more socially competent than children
exposed to other forms of maltreatment.
This may reflect a lack of chronicity associated
with sexual abuse, or it may suggest that problems
related to sexual abuse may not reveal themselves
until later periods in a child’s development, when
issues concerning sexuality become more salient.
Other studies have found that psychological maltreatment
and emotional abuse account for most
of the distortions in development attributed to
maltreatment in general, and have the most
negative consequences for a child (Crittenden,
Claussen, & Sugarman, 1994).
The example of maltreatment illustrates how
contexts for development encompass heterogeneous
sets of circumstances, and how child outcomes
may vary as a function of (1) the configuration
of these circumstances over time, (2) when
and where outcomes are assessed, and (3) the
specific aspects of development that are affected.
More precise definitions are needed if
the impact of maltreatment, or for that matter
any contextual event (e.g., parent disciplinary
styles, family support, intellectual stimulation),
is to be understood.
Even for those forms of child psychopathology
for which there are strong neurobiological
influences, the expression of the disorder is
likely to interact with contextual demands. For
example, Iaboni, Douglas, and Baker (1995)
found that although the overall pattern of responding
shown by children with ADHD was
indicative of a generalized inhibitory deficit, the
self-regulatory problems of these children became
more evident with continuing task demands
for inhibition and/or deployment of effort.
Likewise, tasks having high interest value
or high external incentives may moderate these
children’s typically deficient performance on
less interesting or low incentive tasks (Carlson
& Tamm, 2000; Slusarek, Velling, Bunk, &
Eggers, 2001).
Child psychopathology research has increasingly
focused on the role of the family system,
the complex relationships within families, and
the reciprocal influences among various family
subsystems (Fiese, Wilder, & Bickham, 2000).
There is a need to consider not only the processes
occurring within disturbed families, but
the common and unique ways in which these
processes affect both individual family members
and subsystems. Within the family, the roles of
the mother–child and marital/couple subsystems
have received the most research attention to
date, with less attention given to the roles of siblings
(Hetherington, Reiss, & Plomin, 1994) and
fathers (Lamb & Billings, 1997; Phares &
Compas, 1992). For the most part, research into
family processes and child psychopathology has
not kept pace with family theory and practice,
and there is a need for the development of sophisticated
methodologies and valid measures
that will capture the complex relationships hypothesized
to be operative in disturbed and
normal family systems (Bray, 1995; Bray, Maxwell,
& Cole, 1995). This task is complicated by
a lack of consensus concerning how dysfunctional
or healthy family functioning should be
defined, what specific family processes are important
to assess (Bray, 1994; Mash & Johnston,
1995), or the extent to which such measures of
family environment reflect true environmental
effect or shared genetic influences between
parent and child (Plomin, 1995).
DEFINING CHILD
PSYCHOPATHOLOGY
There has been, and continues to be, a lack of
consensus concerning how psychopathology in
children should be defined (Silk et al., 2000;
Sonuga-Barke, 1998). Although the situation
is improving, comparisons of findings across
studies are extremely difficult to make, because
of the idiosyncratic ways in which samples of
children have been constituted. For example,
children described as “hyperactive” in previous
studies have varied widely with respect to their
symptoms and conditions, problem severities,
comorbidities, and levels of cognitive functioning.
More recently, researchers and clinicians have
come to define child psychopathology using stan1.
A Developmental–Systems Perspective 25
dardized diagnostic systems such as DSM-IV
(American Psychiatric Association [APA], 1994,
2000) and the International Classification of Diseases,
10th revision (ICD-10; WHO, 1992). The
diagnostic criteria utilized in DSM-IV are the
ones most commonly used in North America, and
these are presented for the individual disorders
described in each of the chapters of this volume.
However, the increased use and acceptance of
DSM-IV should not be taken as an indication of
widespread agreement regarding the fundamental
nature of what constitutes psychopathology in
children or the specific criteria used to define it
(cf. Achenbach, 1997; Cantwell, 1996; Follette &
Houts, 1996; Scotti, Morris, McNeil, & Hawkins,
1996). In many ways, the increased use of DSMIV
seems to reflect a degree of resignation on the
part of researchers and clinicians concerning the
prospects for developing a widely agreed-upon
alternative approach, combined with a growing
consensus regarding the need to achieve a greater
level of standardization (albeit an imperfect one)
in defining childhood disorders.
Several fundamental questions have characterized
most discussions concerning how child psychopathology
should be defined:
1. Should child psychopathology be viewed as
a disorder that occurs within the individual
child, as a relational disturbance, as a reaction
to environmental circumstances, or as some
combination of all of these?
2. Does child psychopathology constitute a condition
qualitatively different from normality (aberration),
an extreme point on a continuous trait
or dimension, a delay in the rate at which a
normal trait would typically emerge, or some
combination of the three? How are “subthreshold”
problems to be handled?
3. Can homogeneous disorders be identified, or
is child psychopathology best defined as a
configuration of co-occurring disorders or as
a profile of traits and characteristics?
4. Can child psychopathology be defined as a
static entity at a particular point in time, or do
the realities of development necessitate that
it be defined as a dynamic and ongoing process
that expresses itself in different ways over
time and across contexts?
5. Is child psychopathology best defined in terms
of its current expression, or do definitions also
need to incorporate nonpathological conditions
that may constitute risk factors for later
problems?
There are currently no definitive answers to
these questions. More often, the way in which
they are answered reflects theoretical or disciplinary
preferences and specific purposes and goals
(e.g., defining samples for research studies, or
determining program or insurance eligibility).
Psychopathology as
Adaptational Difficulty
As we have noted earlier, a common theme in
defining child psychopathology has been that of
adaptational difficulty or failure (Garber, 1984;
Mash, 1998). Sroufe and Rutter (1984) note that
regardless of whether “particular patterns of early
adaptation are to a greater or lesser extent influenced
by inherent dispositions or by early experience,
they are nonetheless patterns of adaptation”
(p. 23). Developmental competence is
reflected in a child’s ability to use internal and
external resources to achieve a successful adaptation
(Masten & Curtis, 2000; Waters & Sroufe,
1983), and problems occur when the child fails
to adapt successfully. Even with wide variations
in terminology and proposed explanatory mechanisms
across theories, there is general agreement
that maladaptation represents a pause, a regression,
or a deviation in development (Garber,
1984; Simeonsson & Rosenthal, 1992).
In conceptualizing and defining psychopathology
as adaptational difficulty, it is also essential
to conceptualize and identify the specific
developmental tasks that are important for children
at various ages and periods of development,
and the many contextual variables that
derive from and surround the child (Garber,
1984; Luthar et al., 1997; Mash, 1998). In this
regard, the study of psychopathology in children
and the study of development and context
are for all intents and purposes inseparable
(Cicchetti & Aber, 1998).
In determining whether a given behavior
should be considered to be deviant in relation to
stage-salient developmental issues, Garber (1984)
stresses the need to understand several important
parameters. The first, “intensity,” refers to the
magnitude of behavior as excessive or deficient.
The second, “frequency,” refers to the severity
of the problem behavior, or how often it does
or does not occur. Third, the “duration” of behavior
must be considered. Some difficulties are
transient and spontaneously remit, whereas
others persist over time. To these parameters, we
would add a qualitative parameter reflecting how
26 I. INTRODUCTION
grossly atypical the behavior may be (e.g., some
of the complex compulsions seen in Tourette’s
disorders), such that even low-intensity, lowfrequency,
and short-duration behavior may be
so bizarre as to constitute “psychopathology.” It
is crucial that the intensity, frequency, duration,
and atypicality of the child’s behavior be appraised
with respect to what is considered normative for a
given age. The final parameter of deviance concerns
the “number of different symptoms” and
their “configuration.” Each of these parameters is
central to research and theory, and to one’s specific
definition of adaptational failure, regression,
stagnation, or deviation.
Social Judgment
The diagnosis of psychopathology in children is
almost always a reflection of both the characteristics
and behavior of the child and of significant
adults and professionals (Lewis, 2000). Research
findings utilizing behavior problem checklists and
interviews indicate that there can be considerable
disagreement across informants (e.g., parents,
teachers, professionals) concerning problem behaviors
in children (Achenbach, McConaughy, &
Howell, 1987; Feiring & Lewis, 1996). Mothers
typically report more problems than do fathers
(e.g., Achenbach et al., 1991), and across a range
of domains, teachers identify more problems than
other informants do in assessing the same domains.
For example, in a study with maltreated
children, only 21% of children were classified as
resilient by teachers, whereas 64% of children
were so classified based on reports from other
sources (Kaufman et al., 1994).
Issues regarding disagreement–agreement
among informants are complicated by the fact
that the amount of agreement will vary with the
age and sex of the child (Offord et al., 1989), the
nature of the problem being reported on (e.g.,
internalizing vs. externalizing), the method used
to gather information (e.g., interview vs. questionnaire),
and the informants being compared. For
example, Tarullo, Richardson, Radke-Yarrow,
and Martinez (1995) found that both mother–
child and father–child agreement was higher for
preadolescent than for adolescent children and,
in a meta-analysis, Duhig, Renk, Epstein, and
Phares (2000) reported higher mother–father
agreement for externalizing than for internalizing
problems. Disagreements among informants
create methodological difficulties in interpreting
epidemiological data when such data are obtained
from different sources, and also in how
specific diagnoses are arrived at in research and
practice.
Also of importance is how disagreements among
informants are interpreted. For example, disagreements
may be viewed as (1) reflections of bias or
error on the part of one informant; (2) evidence
for the variability of children’s behavior across the
situations in which they are observed by others; (3)
lack of access to certain types of behavior (i.e.,
private events) on the part of one informant; (4)
denial of the problem; or (5) active distortion of
information in the service of some other goal (e.g.,
defensive exclusion, treatment eligibility).
Parental psychopathology may “color” descriptions
of child problems—as may occur when abusive
or depressed mothers provide negative or
exaggerated descriptions of their children (Gotlib
& Hammen, 1992; Mash et al., 1983; Richters,
1992), or when dismissive/avoidant adult informants
deny the presence of emotional problems
at the same time that professionals observe a high
level of symptoms (Dozier & Lee, 1995). These
latter types of problems in reporting may be especially
likely, given the frequent lack of correspondence
between the expression and the experience
of distress for many child and adult disturbances.
Hypothesized relationships between parental
psychopathology and reports of exaggerated child
symptoms have received mixed support. For example,
some studies have failed to find evidence
for distorted reports by depressed mothers
(Tarullo et al., 1995).
TYPES OF CHILD
PSYCHOPATHOLOGY
The types of problems for which children are referred
for treatment are reflected in the different
approaches that have been used to conceptualize
and classify these problems. Among the more common
of these approaches are the following:
1. General and specific behavior problem checklists,
which enumerate individual child symptoms—
for example, the Child Behavior Checklist
(Achenbach, 1991) and the Children’s
Depression Inventory (Kovacs & Beck, 1977).
2. Dimensional approaches, which focus on
symptom clusters or syndromes derived from
behavior problem checklists—for example,
the Child Behavior Checklist and Profile
(Achenbach, 1993).
1. A Developmental–Systems Perspective 27
3. Categorical approaches, which use predetermined
diagnostic criteria to define the presence
or absence of particular disorders—for
example, the DSM-IV (APA, 1994) and ICD-
10 (WHO, 1992).
4. A multiple-pathway, developmental approach,
which emphasizes developmental antecedents
and competencies both within the
child and the environment that contribute to
(mal)adjustment and (mal)adaptation (Sroufe,
1997).
Issues related to the use of these different
classification approaches are discussed in a later
section of this chapter. What follows is a brief
overview of the types of problem behaviors,
dimensions, and disorders that occur during
childhood and that are the topics of this volume’s
other chapters.
Individual Symptoms
The individual behavioral and emotional problems
(i.e., symptoms) that characterize most
forms of child psychopathology have been found
to occur in almost all children at one time or another
during their development (e.g., Achenbach
& Edelbrock, 1981; Achenbach et al., 1991;
MacFarlane et al., 1954). When taken in isolation,
specific symptoms have generally shown little
correspondence to a child’s overall current adjustment
or to later outcomes. This is the case even
for many symptoms previously hypothesized to
be significant indicators of psychopathology
in children—for example, thumbsucking after
4 years of age (Friman, Larzelere, & Finney,
1994). Usually the age-appropriateness, clustering,
and patterning of symptoms are what serve
to define child psychopathology, rather than the
presence of individual symptoms.
Many of the individual behavior problems displayed
by children referred for treatment are
similar to those that occur in less extreme forms
in the general population or in children of
younger ages. For example, Achenbach et al.
(1991) found that although referred children
scored higher than nonreferred children on 209
of 216 parent-rated problems, only 9 of the 209
items showed effects related to clinical status that
were considered to be large (accounting for more
than 13.8% of the variance), according to criteria
specified by Cohen (1988). To illustrate the
kinds of individual symptoms that are more common
in referred than in nonreferred children,
individual parent-reported symptoms that accounted
for 10% or more of the variance in clinical
status in the Achenbach et al. (1991) study are
shown in Table 1.1. It can be seen that even the
problems that best discriminated between referred
and nonreferred children are relatively
common behaviors that occur to some extent in
all children—they are not particularly strange or
unusual behaviors. In addition, most individual
problem behaviors (approximately 90% of those
on behavior problem checklists) do not, by themselves,
discriminate between groups of clinicreferred
and nonreferred children. Nondiscriminating
items include some problems for children
in both groups that are relatively common (e.g.,
“brags,” “screams”) and others that occur less frequently
(e.g., “sets fires,” “bowel movements outside
the toilet”).
Dimensions of Child Psychopathology
A second approach to describing child psychopathology
identifies symptom clusters or “syndromes”
derived through the use of multivariate
statistical procedures, such as factor analysis
or cluster analysis (e.g., Achenbach, 1993, 1997;
McDermott, 1993; McDermott & Weiss, 1995).
Research has identified two broad dimensions of
child psychopathology—one reflecting “externalizing”
or “undercontrolled” problems, and the
other reflecting “internalizing” or “overcontrolled”
problems (Reynolds, 1992). The externalizing
dimension encompasses behaviors
often thought of as directed at others, whereas
the internalizing dimension describes feelings
or states that are commonly viewed as “innerdirected.”
Within the two broad dimensions of externalizing
and internalizing disorders are specific subdimensions
or syndromes. Some subdimensions
of child psychopathology that have commonly
been identified in research are presented in
Table 1.2. They include “withdrawn,” “somatic
complaints,” “anxious/depressed,” “social problems,”
“thought problems,” “attention problems,”
“delinquent behavior,” and “aggressive behavior”
(Achenbach, 1993). Examples of the specific
problem behaviors constituting each of these
subdimensions are also included in Table 1.2. The
particular subdimensions that are identified may
vary from study to study as a function of the item
pool from which they are derived, the age and sex
of children in the sample, the methods of assessment,
and the informants.
28 I. INTRODUCTION
Taxometric efforts have also described groups
of children in terms of consistently identified
profiles of scores on the various syndromes
(Achenbach, 1993). Such profiles have been reliably
identified and appear to have promise in
addressing problems related to comorbidity (see
the section on comorbidity, below). At present,
however, our nomenclature for describing these
profiles is limited, and they have yet to be widely
validated or used in clinical research and practice.
Categories of Child Psychopathology
The DSM-IV diagnostic system (APA, 1994,
2000) provides comprehensive coverage of the
general types of symptom clusters displayed by
children characterized as having mental disorders.
To illustrate, DSM-IV categories that apply
to children are listed in Tables 1.3 to 1.6. These
tables are not intended to be exhaustive of all
DSM-IV diagnoses that may apply to children.
Rather, they are intended to provide an overview
of the range and variety of disorders that typically
occur during childhood. Specific DSM-IV disorders
and their subtypes are discussed in detail in
the subsequent chapters of this volume.
Table 1.3 lists the DSM-IV categories for developmental
and learning disorders, including
mental retardation, pervasive developmental disorders
(e.g., autistic disorder), specific problems
related to reading and mathematics, and communication
difficulties. Many of these disorders
constitute chronic conditions that often reflect
deficits in capacity rather than performance difficulties
per se.
Table 1.4 lists DSM-IV categories for other
disorders that are usually first diagnosed in infancy,
childhood, or adolescence. These disorders
have traditionally been thought of as first occurring
in childhood or as exclusive to childhood and
as requiring operational criteria different from
those used to define disorders in adults.
Table 1.5 lists disorders that can be diagnosed
in children or adolescents (e.g., mood disorders,
anxiety disorders), but that are not listed in DSMIV
as distinct disorders first occurring during
childhood, or requiring operational criteria that
are different from those used for adults. In many
ways, the DSM-IV distinction between child and
adult categories is an arbitrary one; it is more a
reflection of our current lack of knowledge concerning
the continuities between child and adult
disorders than of the existence of qualitatively
distinct conditions. Recent efforts to diagnose
ADHD in adults illustrate this problem. Although
the criteria for ADHD were derived from work
with children, and the disorder is included in the
“infancy, childhood, or adolescence” section of
DSM-IV, these criteria are being used to diagnose
adults even though they do not fit the expression
of the disorder in adults very well.
The more general issue here is whether there
is a need for separate diagnostic criteria for children
versus adults, or whether one can use the
same criteria by adjusting them to take into account
differences in developmental level. For
instance, the childhood category of overanxious
disorder in DSM-III-R (APA, 1987) was subsumed
under the category of generalized anxiety
disorder in DSM-IV (APA, 1994). With this
TABLE 1.1. Individual Parent-Rated Problems
Accounting for More than 10% of the Variance in
Clinical Status of Children Aged 4–16
Poor school work (19%)a,b
Can’t concentrate, can’t pay attention for long (18%)b
Lacks self-confidence (17%)b
Punishment doesn’t change his/her behavior (17%)b
Disobedient at home (15%)b
Has trouble following directions (15%)b
Sad or depressed (15%)b
Uncooperative (14%)b
Nervous, high-strung, or tense (14%)b
Feels he/she can’t succeed (13%)
Feels worthless or inferior (13%)
Disobedient at school (13%)
Easily distracted (13%)
Lies (13%)
Looks unhappy without good reason (13%)
Fails to finish things he/she starts (12%)
Defiant (12%)
Doesn’t get along with other kids (12%)
Has a hard time making friends (12%)
Doesn’t seem to feel guilty after misbehavior (12%)
Needs constant supervision (12%)
Sudden changes in mood or feelings (12%)
Angry moods (11%)
Impulsive or acts without thinking (11%)
Irritable (11%)
Temper tantrums or hot temper (10%)
Does things slowly and incorrectly (10%)
Loses train of thought (10%)
Loss of ability to have fun (10%)
Passive or lacks initiative (10%)
Note. Data from Achenbach, Howell, Quay, and Conners (1991,
pp. 107–115).
aNumber in parentheses indicates the percentage of variance
accounted for by this problem behavior.
bItems accounting for 14% or more of the variance are designated
as having a large effect size, according to criteria presented
by Cohen (1988).
1. A Developmental–Systems Perspective 29
TABLE 1.2. Commonly Identified Dimensions of Child Psychopathology and Examples of
Items Reflecting Each of the Dimensions
Withdrawn
Would rather be alone
Refuses to talk
Secretive
Shy, timid
Stares blankly
Sulks
Underactive
Unhappy, sad, depressed
Withdrawn
Somatic complaints
Feels dizzy
Overtired
Aches, pains
Headaches
Nausea
Eye problems
Rashes, skin problems
Stomachaches
Vomiting
Anxious/depressed
Lonely
Cries a lot
Fears impulses
Needs to be perfect
Feels unloved
Feels persecuted
Feels worthless
Nervous, tense
Fearful, anxious
Feels too guilty
Self-conscious
Suspicious
Unhappy, sad, depressed
Worries
Harms self
Thinks about suicide
Overconforms
Hurt when criticized
Anxious to please
Afraid of mistakes
Social problems
Acts too young
Too dependent
Doesn’t get along with peers
Gets teased
Not liked by peers
Clumsy
Prefers younger children
Overweight
Withdrawn
Lonely
Cries
Feels unloved
Feels persecuted
Feels worthless
Accident-prone
Thought problems
Can’t get mind off thoughts
Hears things
Repeats acts
Sees things
Strange behavior
Strange ideas
Stares blankly
Harms self
Fears
Stores up things
Attention problems
Acts too young
Can’t concentrate
Can’t sit still
Confused
Daydreams
Impulsive
Nervous, tense
Poor school work
Clumsy
Stares blankly
Twitches
Hums, odd noises
Fails to finish
Fidgets
Difficulty with directions
Difficulty learning
Apathetic
Messy work
Inattentive
Underachieving
Fails to carry out tasks
Delinquent behavior
Lacks guilt
Bad companions
Lies
Prefers older kids
Runs away from home
Sets fires
Steals at home
Swearing, obscenity
Truancy
Alcohol, drugs
Thinks about sex too much
Vandalism
Tardy
Aggressive behavior
Argues
Brags
Mean to others
Demands attention
Destroys own things
Destroys others’ things
Disobedient at school
Jealous
Fights
Attacks people
Screams
Shows off
Stubborn, irritable
Sudden mood changes
Talks too much
Teases
Temper tantrums
Threatens
Loud
Disobedient at home
Defiant
Disturbs others
Talks out of turn
Disrupts class
Explosive
Easily frustrated
Note. Dimensions are based on analyses across informants (e.g., parents, teachers, and children)
and assessment methods (Child Behavior Checklist, Youth Self-Report Form, and Teacher Report
Form). Adapted from Achenbach (1993, pp. 41–43). Copyright 1993 by T. M. Achenbach. Adapted by
permission.
30 I. INTRODUCTION
TABLE 1.3. DSM-IV Categories for Developmental
and Learning Disorders Usually First Diagnosed in
Infancy, Childhood, or Adolescence
Mental retardation
Mild, moderate, severe, profound, severity unspecified
Learning disorders
Reading disorder
Mathematics disorder
Disorder of written expression
Learning disorder not otherwise specified
Motor skills disorder
Developmental coordination disorder
Communication disorders
Expressive language disorder
Mixed receptive–expressive language disorder
Phonological disorder
Stuttering
Communication disorder not otherwise specified
Pervasive developmental disorders
Autistic disorder
Rett’s disorder
Childhood disintegrative disorder
Asperger’s disorder
Pervasive developmental disorder not otherwise
specified
change, the number of criteria required for children
to meet this diagnosis was also altered.
Finally, Table 1.6 lists DSM-IV categories for
other conditions that are not defined as mental
disorders, but that may be a focus of clinical attention
during childhood or adolescence. The
categories that are included are the ones that
seem especially relevant to children, in that they
emphasize relational problems, maltreatment,
and academic and adjustment difficulties.
APPROACHES TO
THE CLASSIFICATION
AND DIAGNOSIS OF CHILD
PSYCHOPATHOLOGY
The formal and informal classification systems
that have been used by psychiatrists, psychologists,
and educators to categorize the different
forms of child psychopathology have played a
central role in defining the field. For example, in
referring to these systems, Adelman (1995) states:
“They determine the ways individuals are described,
studied, and served; they shape prevailing
practices related to intervention, professional
TABLE 1.4. DSM-IV Categories for Other Disorders
Usually First Diagnosed in Infancy, Childhood,
or Adolescence
Attention-deficit and disruptive behavior disorders
Attention-deficit/hyperactivity disorder
Predominantly inattentive type
Predominantly hyperactive–impulsive type
Combined type
Attention-deficit/hyperactivity disorder not
otherwise specified
Disruptive behavior disorders
Conduct disorder
Oppositional defiant disorder
Disruptive behavior disorder not otherwise specified
Feeding and eating disorders of infancy or early
childhood
Pica
Rumination disorder
Feeding disorder of infancy or early childhood
Tic disorders
Tourette’s disorder
Chronic motor or vocal tic disorder
Tic disorder not otherwise specified
Elimination disorders
Encopresis
Enuresis
Other disorders of infancy, childhood, or adolescence
Separation anxiety disorder
Selective mutism
Reactive attachment disorder of infancy or early
childhood
Stereotypic movement disorder
Disorder of infancy, childhood, or adolescence not
otherwise specified
training, and certification; and they influence
decisions about funding. It is not surprising,
therefore, that debates about classification
schemes, specific diagnostic procedures, and the
very act of labeling are so heated” (p. 29).
Although early conceptualizations of psychopathology
included underdeveloped and global
descriptions of childhood disorders (e.g., “adjustment
problem”), this state of affairs has been
steadily improving. Nevertheless, problems and
issues in describing and classifying childhood disorders
continue to plague the field (e.g., Quay,
Routh, & Shapiro, 1987). As noted by Rutter and
Garmezy (1983), “All too frequently findings have
been inconclusive because the measures employed
have been weak, nondiscriminating, or
open to systematic bias. Similarly, comparisons
1. A Developmental–Systems Perspective 31
between studies have often been vitiated because
cases have been defined differently, because the
settings have been noncomparable, or because
the measures focused on different aspects of behavior”
(p. 865).
There is general agreement in medicine, psychiatry,
and psychology regarding the need for a
system of classifying for childhood disorders.
However, major areas of contention have arisen
around such issues as which disorders should be
included in the system, what the optimal strategies
are for organizing and grouping disorders,
and what specific criteria should be used to define
a particular disorder (Achenbach, 1985;
Achenbach & Edelbrock, 1989; Mash & Terdal,
1997a; Sonuga-Barke, 1998).
The two most common approaches to the diagnosis
and classification of child psychopathology
involve the use of (1) “categorical” classification
systems that are based primarily on informed
clinical consensus, an approach that has dominated
and continues to dominate the field (APA,
1994, 2000); and (2) empirically based “dimensional”
classification schemes derived through the
use of multivariate statistical techniques (Achenbach,
1993, 1997). In addition, alternative and/
or derivative approaches to classification have
been proposed to address perceived deficiencies
associated with the use of categorical and dimensional
approaches. These have included developmentally
based measures (Garber, 1984; Mohr &
Regan-Kubinski, 1999; Sroufe, 1997); laboratory
and performance-based measures (Frick, 2000);
prototype classification (Cantor et al., 1980); and
behavioral classification based on behavioral
excesses, deficits, and faulty stimulus control
(Adams, Doster, & Calhoun, 1977; Kanfer &
Saslow, 1969; Mash & Hunsley, 1990). Although
each of these alternative approaches has something
to offer to the classification of childhood
disorders, they are generally underdeveloped and
unstandardized, and have not been widely accepted
or used in either research or practice.
To date, no single classification scheme for
childhood disorders has established adequate
reliability and validity (Cantwell, 1996; Mash &
Terdal, 1997a). Many researchers and clinicians
continue to express concerns that current diagnostic
and classification systems (1) underrepresent
disorders of infancy and childhood;
(2) are inadequate in representing the interrelationships
and overlap that exist among many
childhood disorders; (3) are not sufficiently sensitive
to the developmental, contextual, and relational
parameters that are known to characterize
most forms of psychopathology in children; and
(4) are heterogeneous with respect to etiology
(Jensen & Hoagwood, 1997; Kagan, 1997).
TABLE 1.5. Selected Categories for Disorders of
Childhood or Adolescence That Are Not Listed
Separately in DSM-IV as Those Usually First
Diagnosed in Infancy, Childhood, or Adolescence
Mood disorders
Depressive disorders
Major depressive disorder
Dysthymic disorder
Bipolar disorders
Anxiety disorders
Specific phobia, social phobia, obsessive–compulsive
disorder, posttraumatic stress disorder, acute stress
disorder, generalized anxiety disorder, anxiety
disorder due to . . . (specific medical condition)
Somatoform disorders
Factitious disorders
Dissociative disorders
Sexual and gender identity disorders
Eating disorders
Sleep disorders
Schizophrenia and other psychotic disorders
Substance-related disorders
Impulse-control disorders not elsewhere classified
Adjustment disorders
Personality disorders
TABLE 1.6. Selected DSM-IV Categories for Other
Conditions That May Be a Focus of Clinical Attention
during Childhood or Adolescence, but Are Not
Defined as Mental Disorders
Relational problems
Relational problem related to a general mental
disorder or general medical condition
Parent–child relational problem
Partner relational problem
Sibling relational problem
Relational problem not otherwise specified
Problems related to abuse or neglect
Physical abuse of child
Sexual abuse of child
Neglect of child
Bereavement
Borderline intellectual functioning
Academic problem
Child or adolescent antisocial behavior
Identity problem
32 I. INTRODUCTION
Categorical Approaches
Categorical approaches to the classification of
childhood disorders have included systems developed
by the Group for the Advancement of Psychiatry
(1974), the WHO (1992), the APA (1994),
and the Zero to Three/National Center for Clinical
Infant Programs (1994). Although a detailed
review of all these systems is beyond the scope
of this chapter, a brief history of the APA’s development
of the DSM approach is presented to
illustrate the issues associated with categorical
approaches, the growing concern for more reliable
classification schemes for childhood disorders,
and the evolving conceptualizations of childhood
disorders over the past 50 years. Also, the
Diagnostic Classification of Mental Health and
Developmental Disorders of Infancy and Early
Childhood, or Diagnostic Classification: 0–3
(DC:0–3; Zero to Three/National Center for
Clinical Infant Programs, 1994), is described to
illustrate a categorical approach that attempts to
integrate developmental and contextual information
into the diagnosis of infants’ and young
children’s problems.
Development of the DSM Approach
One of the first efforts to collect data on mental
illness was in the U.S. census of 1840, which recorded
the frequency of a single category of
“idiocy/insanity.” Forty years later, seven categories
of mental illness were identified: dementia,
dipsomania, epilepsy, mania, melancholia,
monomania, and paresis (APA, 1994). Much later
(in the 1940s), the WHO classification system
emerged with the manuals of the ICD, whose 6th
revision included, for the first time, a section for
mental disorders (APA, 1994; Cantwell, 1996).
In response to perceived inadequacies of the
ICD system for classifying mental disorders, the
APA’s Committee on Nomenclature and Statistics
developed the DSM-I in 1952 (APA, 1952).
There were three major categories of dysfunction
in the DSM-I—“organic brain syndromes,” “functional
disorder,” and “mental deficiency” (Kessler,
1971)—under which were subsumed 106 categories
(by contrast, DSM-IV consists of 407 separate
categories; Cantwell, 1996). The term “reaction”
was used throughout the text, which
reflected Adolf Meyer’s psychobiological view
that mental illness involves reactions of the personality
to psychological, social, and biological
factors (APA, 1987). Children were virtually
neglected in the early versions of DSM, with most
childhood disorders relegated to the adult categories
(Cass & Thomas, 1979; Silk et al., 2000).
In fact, DSM-I included only one child category
of “adjustment reactions of childhood and of adolescence,”
which was included under the heading
of “transient situational disorders.”
As reflected in the use of the term “reaction,”
psychoanalytic theory had a substantial influence
on the classification of both child and adult psychopathology
(Clementz & Iacono, 1993). In
part, this was due to the fact that the first classification
system to focus on childhood psychopathology
was developed by Anna Freud in 1965
(see Cantwell, 1996). Although the term “reaction”
was eliminated from DSM-II (APA, 1968),
a separate section was reserved for classifying
neuroses, and diagnoses could be based on either
an assessment of the client’s presenting symptomatology
or inferences about his or her unconscious
processes (Clementz & Iacono, 1993).
Once again, apart from conditions subsumed
under the adult categories, DSM-II gave little
recognition to childhood difficulties except for
mental retardation and schizophrenia—childhood
type (Cass & Thomas, 1979).
As a formal taxonomy, DSM-III (APA, 1980)
represented a significant advance over the earlier
editions of the DSM. The first and second
editions contained only narrative descriptions of
symptoms, and clinicians had to draw on their
own definitions for making a diagnosis (APA,
1980). In DSM-III, these descriptions were replaced
by explicit criteria, which in turn enhanced
diagnostic reliability (Achenbach, 1985; APA,
1980). Moreover, unsubstantiated inferences that
were heavily embedded in psychoanalytic theory
were dropped; more child categories were included;
a multiaxial system was adopted; and a
greater emphasis was placed on empirical data
(Achenbach, 1985). These changes reflected the
beginnings of a conceptual shift in both diagnostic
systems and etiological models away from an
isolated focus of psychopathology as existing
within the child alone, and toward an increased
emphasis on his or her surrounding context.
DSM-III was revised in 1987 (DSM-III-R) to
help clarify the numerous inconsistencies and
ambiguities that were noted in its use. For example,
empirical data at that time did not support
the category of attention deficit disorder
without hyperactivity as a unique symptom cluster
(Routh, 1990), and this category was removed
from DSM-III-R. DSM-III-R was also developed
1. A Developmental–Systems Perspective 33
to be polythetic, in that a child could be diagnosed
with a certain subset of symptoms without having
to meet all criteria. This was an important change,
especially in light of the heterogeneity and rapidly
changing nature of most childhood disorders
(Mash & Terdal, 1997a). Relative to its predecessors,
far greater emphasis was also placed on
empirical findings in the development of the
DSM-IV, particularly for the child categories.
In order to bridge the planned 12-year span
between the DSM-IV and DSM-V, a revision
(DSM-IV-TR) of the DSM was published in 2000
(APA, 2000). The DSM-IV-TR was limited to text
revisions (e.g., associated features and disorders,
prevalence) and was designed mainly to correct
any factual errors in DSM-IV, make sure that
information is still current, and incorporate new
information since the time the original DSM-IV
literature reviews were completed in 1992. Substantive
changes in diagnostic criteria were not
considered or made; nor were there any changes
in relation to new disorders or subtypes. Thus
DSM-IV and DSM-IV-TR are equivalent with
respect to specific diagnostic criteria.
DSM-IV is a multiaxial system that includes
five different axes. Axis I is used to report clinical
disorders and other conditions that may be a
focus of clinical attention. The various Axis I diagnostic
categories that apply to infants, children,
and adolescents have been listed in Tables 1.3 to
1.6 of this chapter. Axis II includes personality
disorders and mental retardation. The remaining
axes pertain to general medical conditions (Axis
III), psychosocial and environmental problems
(Axis IV), and global assessment of functioning
(Axis V).
Although DSM-III-R (APA, 1987) and DSMIV
(APA, 1994) include numerous improvements
over the previous DSMs—with their greater
emphasis on empirical research, and more explicit
diagnostic criteria sets and algorithms—
criticisms have also been raised (e.g., Mohr &
Regan-Kubinski, 1999; Nathan & Lagenbucher,
1999; Sonuga-Barke, 1998; Sroufe, 1997). One
major criticism is the static nature of DSM categories,
especially when one considers the dynamic
nature of development in children (Mash
& Terdal, 1997a; Routh, 1990). Another source
of dissatisfaction is that the DSM-IV categorical
scheme may contribute minimally to meeting
children’s needs. For example, it may be necessary
for a child to meet specific diagnostic criteria
for a learning disability in order to qualify for
a special education class. However, if the child’s
problems are subclinical, or the child’s problems
relate to more than one DSM category, then he
or she may be denied services (Achenbach, 2000).
However, even if one were to adopt a more dimensional
approach to classification, there would
nonetheless continue to be a categorical interpretation
of the data (e.g., distinguishing between
individuals who require help and those who do
not) (Sonuga-Barke, 1998).
Another problem with DSM-IV relates to the
wording and the lack of empirical adequacy for
certain criterion sets. For example, the words
“often” in the criteria for ADHD and conduct
disorder, and “persistent” and “recurrent” in the
criteria for separation anxiety disorder, are not
clearly defined. This ambiguity poses a particular
problem when one considers that the primary
sources of assessment information are often a
child’s parents, whose perception and understanding
of these terms may be idiosyncratic or
inaccurate. This ambiguity and other factors may
contribute to the unreliability or unsuitability of
the DSM for diagnosing certain childhood disorders
(e.g., Nicholls, Chater, & Lask, 2000). A
further difficulty with DSM-IV diagnostic criteria
is the lack of emphasis on the situational or
contextual factors surrounding and contributing
to various disorders. This is a reflection of the fact
that DSM-IV continues to view mental disorder
as individual psychopathology or risk for psychopathology,
rather than in terms of problems in
psychosocial adjustment. One problem with respect
to the atheoretical nature of DSM is that it
has perhaps mistakenly fostered the assumption
that a description of symptoms is sufficient for
diagnosis, without taking into account natural
history, psychosocial correlates, biological factors,
or response to treatment (Cantwell, 1996). However,
the consideration in DSM-IV of such factors
as culture, age, and gender associated with
the expression of each disorder is laudable, as
is the increased recognition of the importance
of family problems and extrafamilial relational
difficulties.
The changes in the DSMs from 1952 to 2000
reflect increasing diagnostic accuracy and sophistication.
The transition from “reactive” diagnoses
(DSM-I) and the virtual neglect of childhood
criteria (DSM-I, DSM-II) to an increased number
of child categories, more explicit criteria, and
multiaxial evaluation (DSM-III, DSM-III-R), and
then to an even greater emphasis on empirical
research to guide nomenclature as well as the
increased awareness (and inclusion) of contextual
34 I. INTRODUCTION
and developmental considerations (DSM-IV,
DSM-IV-TR), exemplify important shifts in how
psychopathology in children has come to be conceptualized.
However, along with increased complexity
has come a new set of problems. For example,
the extent to which comorbidity is an
artifact of the DSM’s polythetic criteria or truly
differentiated nosological entities is unclear
(Angold, Costello, & Erkanli, 1999; Nottelmann
& Jensen, 1995), or whether the pendulum has
swung too far from not recognizing psychopathology
in children to identifying and diagnosing too
much (Silk et al., 2000).
It is also the case that ongoing changes in diagnostic
criteria based on new findings and other
considerations (e.g., eligibility for services) are
likely to influence prevalence estimates for many
childhood disorders. For example, current estimates
of autistic disorder are about three times
higher than previous ones (Fombonne, 1999;
Tanguay, 2000); this increase is primarily due to
a broadening of the criteria used to diagnose
autism, as well as increased recognition of milder
forms of the disorder (Bryson & Smith, 1998;
Gillberg & Wing, 1999). There is also ongoing
debate about whether Asperger’s disorder is a
variant of autism or simply describes higherfunctioning
individuals with autism (Schopler,
Mesibov, & Kunce, 1998; Volkmar & Klin, 2000).
The resolution of this debate and prevalence estimates
for both autism and Asperger’s disorder
will depend on how the diagnosis of Asperger’s
disorder is used, since no “official” definition for
this disorder existed until it was introduced in
DSM-IV (Volkmar & Klin, 1998).
Development of the DC:0–3 System
In addition to the limitations noted above, DSMIV
does not provide in-depth coverage of the
mental health and developmental problems of
infants and young children, for whom family relationships
are especially salient. To address this
perceived deficiency, the DC:0–3 was developed
by the Diagnostic Classification Task Force of the
Zero to Three/National Center for Clinical Infant
Programs (Zero to Three/National Center for
Clinical Infant Programs, 1994). DC:0–3 is intended
to provide a comprehensive system for
classifying problems during the first 3–4 years of
life (Greenspan & Wieder, 1994; Lieberman,
Wieder, & Fenichel, 1997). Unlike DSM-IV,
DC:0–3 is based on the explicit premise that diagnosis
must be guided by the principle that all infants
and young children are active participants
in relationships within their families. Hence descriptions
of infant–caregiver interaction patterns,
and of the links between these interaction
patterns and adaptive and maladaptive patterns
of infant and child development, constitute an
essential part of the diagnostic process.
In explicitly recognizing the significance of
relational problems, DC:0–3 includes a relationship
disorder classification as a separate axis (Axis
II) in its multiaxial approach (Axis I, primary diagnosis;
Axis III, medical and developmental disorders
and conditions; Axis IV, psychosocial stressors;
Axis V, functional emotional developmental
level). The diagnosis of relationship disturbances
or disorders is based on observations of parent–
child interaction and the parent’s verbal report
regarding his or her subjective experience of the
child. Relational difficulties are rated with respect
to their intensity, frequency, and duration, and
classified as perturbations, disturbances, or disorders.
In making the DC:0–3 Axis II relationship
disorder diagnosis, three aspects of the relationship
are considered: (1) behavioral quality of the
interaction (e.g., sensitivity or insensitivity in responding
to cues); (2) affective tone (e.g., anxious/
tense, angry); and (3) psychological involvement
(e.g., parents’ perceptions of the child and of what
can be expected in a relationship).
Axis V of DC:0–3, functional emotional development
level, includes the ways in which infants
or young children organize their affective, interactive,
and communicative experiences. Axis V
assessment is based in large part on direct observations
of parent–child interaction. The various
levels include social processes such as mutual
attention, mutual engagement or joint emotional
involvement, reciprocal interaction, and affective/
symbolic communication. Problems may reflect
constrictions in range of affect within levels or
under stress, or failure to reach expected levels
of emotional development.
DC:0–3 is of note in recognizing (1) the significance
of early relational difficulties; (2) the need
to integrate diagnostic and relational approaches
in classifying child psychopathology (Lyons-Ruth,
1995); and (3) the need to apply both quantitative
and qualitative criteria in describing relational
problems. In addition, the dimensions and
specific processes that are used for classification
(e.g., negative affect, unresponsivity, uninvolvement,
lack of mutual engagement, lack of reciprocity
in interaction) include those that have
been identified as important in many develop1.
A Developmental–Systems Perspective 35
mental and clinical research studies on early relationships,
and the system is decidedly more
sensitive to developmental and contextual parameters
than DSM-IV. However, although promising,
DC:0–3 is relatively untested, was generated
on the basis of uncontrolled clinical observations,
is of unknown reliability and validity, and suffers
from many of the same criticisms that have been
noted for DSM-IV (Eppright, Bradley, & Sanfacon,
1998). Nevertheless, the scheme provides
a rich descriptive base for exploring the ways in
which psychopathology is expressed during the
first few years of life, and it calls attention to the
need to examine potential continuities between
early problems and later individual and/or family
disorders (Keren, Feldman, & Tyano, 2001;
Thomas & Clark, 1998; Thomas & Guskin, 2001).
Dimensional Approaches
Dimensional approaches to classification assume
that a number of relatively independent dimensions
or traits of behavior exist, and that all children
possess these to varying degrees. These
traits or dimensions are typically derived through
the use of multivariate statistical methods, such
as factor analysis or cluster analysis (Achenbach,
1993). Empirically derived schemes are more
objective, are potentially more reliable, and allow
for a greater description of multiple symptom
patterns than clinically derived classification systems.
However, there are also a number of problems
associated with their use, including the dependency
of the derived dimensions on sampling,
method, and informant characteristics, and on the
age and sex of the child (Mash & Terdal, 1997a).
As a result, there can be difficulties in integrating
information obtained from different methods,
from different informants, over time, or across
situations. Dimensional approaches have also
shown a lack of sensitivity to contextual influences,
although there have been efforts to develop
dimensional classification schemes based
on item pools that include situational content
(e.g., McDermott, 1993).
The growth in the use of multivariate classification
approaches in child and family assessment
has been fueled by the extensive work of Thomas
Achenbach and his colleagues (see the Achenbach
System of Empirically Based Assessment
[ASEBA]: http://www.ASEBA.org) with the various
parent, teacher, youth, observer, and interview
versions of the Child Behavior Checklist and
Profile (Achenbach, 1993), and by the development
of similar assessment batteries (e.g., the
Behavior Assessment System for Children
[BASC]: Kamphaus et al., 1999; Reynolds &
Kamphaus, 1992). For a comprehensive discussion
of these approaches and the use of empirically
derived classification schemes more generally,
the reader is referred to Achenbach (1985,
1993), Hart and Lahey (1999), and Mash and
Terdal (1997a),
It should also be noted that there has been a
trend toward greater convergence of the categorical
and dimensional approaches to classification.
Many of the items that were retained in
DSM-IV child categories were derived from findings
from multivariate studies, and the process
that led to the development of DSM-IV treated
most childhood disorders as dimensions, albeit
the use of cutoff scores on item lists arbitrarily
created categories out of these dimensions
(Spitzer et al., 1990).
Performance-Based
Diagnostic Information
Performance-based information and/or observational
measures provide additional sources of
diagnostic information that may be sensitive to
differences among children exhibiting similar
self- or other-reported symptoms (Frick, 2000;
Kazdin & Kagan, 1994). These measures assess
children’s performance on standardized tasks,
usually ones that reflect basic biological, cognitive,
affective, or social functioning. For example,
tasks involving behavioral observations of fear and
avoidance, recall memory under stressful conditions,
delayed response times to threatening
stimuli, and the potentiation of the blink reflex
following exposure to a threatening stimulus have
all been suggested as potentially useful in diagnosing
groups and/or subgroups of children with
anxiety disorders (Kazdin & Kagan, 1994; Vasey
& Lonigan, 2000). Similarly, tests of behavioral
inhibition (e.g., the stop-signal paradigm) and
tasks involving sustained attention (e.g., the
continuous-performance test) have proven useful
with children with ADHD (Rapport, Chung,
Shore, Denney, & Isaacs, 2000). Measures of low
resting heart rate as an early biological marker for
later aggressive behavior (Raine, Venables, &
Mednick, 1997); facial emotion recognition tasks
and gambling tasks in identifying children with
psychopathic tendencies (Blair, Colledge, &
Mitchell, 2001; Blair, Colledge, Murray, &
Mitchell, 2001); and a variety of cognitive tasks
36 I. INTRODUCTION
for children with autism (Klinger & Renner,
2000) have also been found to have diagnostic
value.
A study by Rubin, Coplan, Fox, and Calkins
(1995) illustrates the utility of performance-based
diagnostic information. These researchers differentiated
groups of preschool children based on
the two dimensions of “emotionality” (i.e., threshold
and intensity of emotional response) and
“soothability” (i.e., recovery from emotional reaction
based on soothing by self and others), and
on their amount of social interactions with peers.
Children’s dispositional characteristics and behavioral
styles were used to predict outcomes.
Asocial children with poor emotion regulation
had more internalizing problems. In contrast,
social children with poor emotion regulation were
rated as having more externalizing difficulties.
When behavioral and emotional dimensions were
incorporated into classification, it was possible to
make finer predictions—for example, that only a
certain type of asocial children (i.e., reticent children
with poor emotion regulation) would display
later problems.
The use of performance-based measures in
diagnosis is predicated on the availability of reliable
and valid performance indicators for groups
of children with known characteristics. Although
such data are available in varying amounts for a
wide range of disorders, there is a need to validate
such findings for the purposes of diagnosis
and against other sources of information. It is also
the case that performance criteria for these measures
are based on information obtained from
children who were themselves previously identified
using other diagnostic procedures. This raises
the question of nonindependence and representativeness
of data sources. There is also little normative
information available regarding the base
rates of children in the general population who
exhibit certain patterns of responding on these
tasks.
ISSUES IN CLASSIFICATION
Categories, Dimensions, or Both?
Psychological studies of child psychopathology
have tended to conceptualize behavior, affect,
and cognition on quantitative/continuous dimensions,
whereas child psychiatry has tended to
conceptualize child psychopathology in categorical
terms. Both approaches are relevant to classifying
childhood disorders, in that some disorders
may be best conceptualized as qualitatively
distinct conditions and others as extreme points
on one or more continuous dimensions. Kazdin
and Kagan (1994) argue for greater research attention
to qualitatively distinct categories of disorder,
based on illustrative findings from studies
suggesting that the emotional arousal generated
by unfamiliarity, threat, and attack is not a continuous
dimension, and that it is possible to identify
different subgroups of aggressive children
based on varying levels of adrenaline in their
urine.
There is currently little agreement as to which
childhood disorders are best conceptualized as
categories and which as dimensions. It has been
suggested that many childhood disorders, such as
anxiety, depression, ADHD, and the disruptive
behavior disorders, appear to reflect dimensions
of personality rather than categorical problems
(e.g., Werry, 2001). For example, childhood
ADHD symptom clusters of inattention–disorganization
and hyperactivity–impulsivity have been
found to be related to adult personality dimensions
of low conscientiousness and low agreeableness,
respectively (Nigg et al., 2001). Even a disorder
such as autism, which has traditionally been
viewed as “categorical” in nature, can be conceptualized
as an extreme on a continuum of social
behavior (Baron-Cohen, 2000). For dimensional
disorders, children who score just below the cutoff
for a diagnosis may one day meet criteria, and
often show impairment comparable to that of
children who score above the cutoff. Similarly,
those above the cutoff may one day move below
it. Since any classification scheme represents a
construction rather than a reality, it seems unlikely
that most disorders will fall neatly into one
designation or the other (Lilienfeld & Marino,
1995). Whether or not particular conditions are
construed as qualitatively distinct categories, as
continuous dimensions, or as both will probably
depend on the utility, validity, and predictive
value of particular groupings and subgroupings
for certain purposes related to understanding and
remediating child psychopathology. Research
into such subgroupings is just beginning to
emerge (e.g., Kendall, Brady, & Verduin, 2001).
Regardless of the particular approach one
adopts for the classification of childhood psychopathology,
diagnostic decisions need to be based
on a comprehensive assessment of the individual
1. A Developmental–Systems Perspective 37
child—one that incorporates sensitivity to and
understanding of the complexity of multiple antecedents,
developmental considerations, comorbidity,
continuity–discontinuity, and the constantly
changing nature of the child (Orvaschel,
Ambrosini, & Rabinovich, 1993).
Comorbidity
An issue that has important ramifications for
theory and research in defining and classifying
child psychopathology is comorbidity (Achenbach,
1995; Angold, Costello, & Erkanli, 1999;
Carey & DiLalla, 1994; Caron & Rutter, 1991;
Sonuga-Barke, 1998). “Comorbidity” generally
refers to the manifestation of two or more disorders
that co-occur more often than would be
expected by chance alone. For example, although
the base rates for ADHD and conduct disorder
in the general population are less than 10% for
each disorder, epidemiological studies have
found that among children diagnosed with
ADHD, approximately 50% are also diagnosed
with conduct disorder (Kazdin & Johnson, 1994;
Loeber & Keenan, 1994). Comorbidity has been
reported to be as high as 50% in community
samples and even higher in clinic samples (Anderson,
Williams, McGee, & Silva, 1987; Bird et al.,
1988; Caron & Rutter, 1991). Some of the more
commonly co-occurring child and adolescent disorders
include conduct disorder and ADHD,
autistic disorder and mental retardation, and
childhood depression and anxiety.
There is continuing debate regarding the definition
and nature of “comorbidity” (Angold,
Costello, & Erkanli, 1999; Blashfield, McElroy,
Pfohl, & Blum, 1994; Caron & Rutter, 1991;
Lilienfeld, Waldman, & Israel, 1994; Meehl,
2001; Robins, 1994; Rutter, 1994b; Sameroff,
2000a; Spitzer, 1994; Widiger & Ford-Black,
1994). Some researchers contend that the term
is wholly inadequate, because it does not distinguish
accurately between manifest conditions
seen in organic medicine (e.g., diseases) and latent
conditions described in mental health (e.g.,
syndromes and disorders (Lilienfeld et al., 1994).
Others argue that the dispute over whether one
should use the term “comorbidity,” “co-occurrence,”
or “covariation” is largely a semantic one
(Rutter, 1994b; Spitzer, 1994; Widiger & Ford-
Black, 1994).
Several possible reasons why comorbidity may
be exaggerated or artificially produced have been
identified in the literature (Angold, Costello, &
Erkanli, 1999; Caron & Rutter, 1991; Lilienfeld
et al., 1994; Rutter, 1994b; Verhulst & van der
Ende, 1993). There may be a sampling bias that
occurs whenever there are fewer numbers of individuals
who are referred to clinics than who
exhibit a given disorder. In such cases, the clinic
samples will contain a disproportionately large
number of subjects who display comorbid conditions.
This phenomenon occurs because the
probability of being referred to mental health
services is higher for a child with a comorbid condition
than for a child with only one disorder.
Related to this sampling bias are various other
referral factors that may inflate the degree of cooccurring
disorders among clinic samples. Clinics
that and clinicians who specialize in treating
more complicated cases, for example, may be more
likely to receive referrals in which comorbid conditions
are present. In addition, children with internalizing
difficulties such as depression are more
likely to be referred by their parents or the school
system if they also show externalizing symptoms,
largely because externalizing problems are viewed
as more disruptive by referral sources.
Comorbidity may also reflect various sources
of nosological confusion arising from the manner
in which different childhood disorders have
been conceptualized and organized. For instance,
Widiger and Ford-Black (1994) claim
that excessive rates of co-occurrence seemed to
appear concomitantly with the changes that occurred
in DSM-III (e.g., increased coverage,
divisions of diagnostic categories, the provision
of separate and multiple axes). Another example
is that DSM-IV makes it possible to have multiple
diagnoses in the absence of multiple syndromes
(Cantwell, 1996; Robins, 1994). One
source of confusion stems from the overlapping
criterion sets within contemporary classification
schemes. In DSM-IV, diagnoses are based on a
set of polythetic criteria that includes specific
symptom constellations. In many cases, the presence
of concomitant symptoms of a different
kind are ignored, resulting in an increased likelihood
that the accompanying symptoms will be
represented in a different diagnostic category
(Caron & Rutter, 1991). Sonuga-Barke (1998)
argues, however, that although earlier diagnostic
systems steered clear of comorbidity by using
a hierarchical set of exclusionary criteria,
“these approaches were abandoned because
they clearly led to a misrepresentation of the
38 I. INTRODUCTION
structure of disorder” (p. 119). For example,
they led to low base rates of disorders and poor
interrater agreement.
Apart from the various artifactual contributors
to comorbidity, there are also indicators in support
of “true” comorbidity (Rutter, 1994b). It is
possible that general propensities toward and/or
struggles with adaptation are at the core of every
disorder, but how the phenotype is expressed is
contingent upon a myriad of environmental conditions
and person–environment interactions
(Caron & Rutter, 1991). Consistent with this
notion, Lilienfeld et al. (1994) maintain that
comorbidity in childhood disorders may be partly
a function of developmental level—that is, of
underlying processes that have not yet achieved
full differentiation. Differing rates of comorbidity
with age may also reflect the fact that the appearance
of one disorder or problem may precede the
appearance of the other, as is the case for anxiety
preceding depression (Brady & Kendall,
1992) or for impulsivity preceding attentional
problems (E. L. Hart et al., 1995). Still another
possibility is that comorbidity reflects “a more
amorphous early expression of psychopathology
in young children that does not crystallize into
more definitive psychopathology until later in
life” (Cantwell, 1996, p. 4). Comorbidity can also
arise as a result of a causal association in which
the severity of one disorder may lead to or greatly
increase the later risk for another disorder (e.g.,
ADHD and oppositional defiant disorder) or a
shared underlying cause, such as common genetic
effects (e.g., conduct disorder and depression) or
shared environmental effects (oppositional defiant
disorder and conduct disorder).
In summary, it would appear that some cases
of comorbidity are the result either of ambiguity
in the definition of dysfunctionality that is used,
or of artifactual/methodological issues. However,
as Kazdin and Kagan (1994) note, “the broader
point is still relevant and not controverted with
specific diagnostic conundrums—namely, multiple
symptoms often go together in packages”
(p. 40). This is not to suggest that all disorders
cluster together into packages; rather, the fact
that many frequently do has important implications
for how child psychopathology is conceptualized
and treated. The complexity of comorbidity
behooves researchers to move beyond
singular models and to examine multiple expressions,
etiologies, and pathways of childhood dysfunction
(Burt, Krueger, McGue, & Iacono,
2001; Kazdin & Johnson, 1994).
THEORY AND CHILD
PSYCHOPATHOLOGY
The Role of Theory in Child
Psychopathology
Every step in the research process is influenced
by the investigator’s preconceptions and ideologies
(Kuhn, 1962; Maxwell & Delaney, 1990). As
the history of child psychopathology has shown,
an overemphasis on a grand theory or explanatory
model in the absence of data can perpetuate
false ideas and seriously impede our understanding
of childhood disorders. On the other
hand, “data gathering in the absence of hypotheses
can become an inconsequential exercise
in gathering inconsequential facts” (Rutter &
Garmezy, 1983, p. 870). The value of theory lies
not just in providing answers but also in raising
new questions, which arise not only from addressing
new problems but also from looking at familiar
problems in different ways. One cannot consider
theory, research, and practice in childhood
psychopathology without also having some understanding
of the underlying philosophical and
epistemological assumptions that have guided
work in this area. In this context, Overton and
Horowitz (1991) discuss four levels of science: (1)
epistemology; (2) guidelines, rules, and definitions
of scientific knowing; (3) metatheoretical
principles; and (4) theory.
The first level, “epistemology,” defined as a
theory about the nature of knowledge itself, has
to do with the general rules of science, the
metatheoretical assumptions about the nature of
humankind, and the specific theoretical models
and research designs that arise out of such assumptions.
One epistemological stance (i.e., “realism”)
asserts that knowledge exists independently
of one’s own perceptual and cognitive
processes (Maxwell & Delaney, 1990; Overton &
Horowitz, 1991). “Logical positivism,” a view that
has guided most of our past and present research
efforts in child psychopathology, reflects this
stance. A second philosophical position is that of
“rationalism.” Rationalists contend that the knower
of scientific knowledge actively constructs what
is known (Maxwell & Delaney, 1990). Instead of
there being a fixed and absolute knowledge base
to unveil, rationalists assume that knowledge
derives from the exercise of relating and interpreting
observables to latent constructs (Overton
& Horowitz, 1991). Within this metatheoretical
position, there lies a continuum between the
1. A Developmental–Systems Perspective 39
belief at one end that our knowledge base will
always be uncertain, and the conviction at the
other end that some universal truth must lie beyond
our interpretive schemes.
At the second level of scientific knowledge—
that of “guidelines, rules, and definitions”—it
becomes evident that epistemology exerts a
strong influence. Logical positivism, for instance,
distinguishes scientific knowledge from knowledge
that accumulates from other modes of
knowing by requiring that all theoretical constructs
be reducible to stable, objective, and observable
knowledge (Maxwell & Delaney, 1990;
Overton & Horowitz, 1991). This view maintains
that theoretical constructs are to be mathematically
related (via correspondence rules) to directly
observable behavior and events. Theory, under
this argument, advances by means of the empirical
method. A hypothesis is tested and when
enough hypotheses have been independently and
empirically supported, generalizations can be
made (via the inductive process) to form a theoretical
model.
At the third level of scientific knowledge identified
by Overton and Horowitz (1991), “metatheoretical
principles” guide the development of
more specific theories. Two metaphors have been
dominant in guiding scientific metatheory: the
“machine” and the “organic” metaphors (Overton
& Horowitz, 1991; Simeonsson & Rosenthal,
1992). The machine metaphor adopts a metatheor