Content uploaded by Filip De Fruyt
Author content
All content in this area was uploaded by Filip De Fruyt on Jul 22, 2015
Content may be subject to copyright.
The Structure of Maladaptive Personality Traits in Childhood: A Step
Toward an Integrative Developmental Perspective for DSM–V
Barbara De Clercq, Filip De Fruyt, Karla Van Leeuwen, and Ivan Mervielde
Ghent University
The present study describes the construction of a taxonomy of trait-related symptoms in childhood, the
Dimensional Personality Symptom Item Pool (DIPSI), and examines the replicability of the taxonomy’s
higher order structure across maternal ratings of referred (N ⫽ 205) and nonreferred (N ⫽ 242) children
and self-ratings of adolescents (N ⫽ 453). The DIPSI’s 4 higher order factors—that is, Emotional
Instability, Disagreeableness, Introversion, and Compulsivity—showed clear correspondence with the
dimensions of personality pathology found in adulthood (Dimensional Assessment of Personality
Pathology—Basic Questionnaire; W. J. Livesley, 1990; Schedule for Nonadaptive and Adaptive Per-
sonality; L. A. Clark, 1993). These 4 factors can be further organized into 2 superfactors, representing
Internalizing and Externalizing Traits, demonstrating empirical and conceptual relationships with psy-
chopathology models in childhood and adulthood. The implications for the assessment and conceptual-
ization of early trait pathology are discussed in the context of an integrative developmental perspective
on the construction of the Diagnostic and Statistical Manual of Mental Disorders—Fifth Edition.
Keywords: childhood, adolescence, personality pathology, psychopathology, DSM–V
The onset of adult psychiatric disorders at a young age has been
demonstrated by different studies, showing important associations
between early behavioral, emotional, or social problems and adult
mental disorders (Cohen et al., 1993; Goodwin, Fergusson, &
Horwood, 2004; Pine, Cohen, Cohen, & Brook, 1999; Roza,
Hofstra, van der Ende, & Verhulst, 2003). However, few studies
have examined childhood antecedents of Axis II personality dis-
orders, because the current edition of the Diagnostic and Statistical
Manual of Mental Disorders (4th ed.; DSM–IV; American Psychi-
atric Association, 1994) primarily postulates the diagnosis of per-
sonality disorders in late adolescence or young adulthood. The
relationships between childhood personality and personality dis-
orders in young adulthood have been described only recently
(Bernstein, Cohen, Skodol, Bezirganian, & Brook, 1996; Kasen,
Cohen, Skodol, Johnson, & Brook, 1999), suggesting that person-
ality disorders do have specific early precursors.
Corroborating these research lines, the present work focuses on
the construction of a taxonomy of trait-related symptoms in child-
hood that are presumed to be precursors of enduring maladaptive
patterns of behavior, thoughts, and feelings characterizing Axis II
personality disorders in adulthood. Such taxonomic work facili-
tates the identification of basic dimensions of child maladaptive
personality functioning and additionally yields a classification of
age-specific symptoms of pathological personality functioning.
Delineating the higher and lower level maladaptive trait structure
at young ages is a necessary step in the study of the developmental
course of personality disorder symptoms across the life span. A
specific child maladaptive trait taxonomy further offers a useful
framework to incorporate a developmental perspective in the con-
struction of the Diagnostic and Statistical Manual of Mental
Disorders—Fifth Edition (DSM–V), a major issue raised by the
DSM–V Research Planning Work Groups (Widiger, Simonsen,
Krueger, Livesley, & Verheul, 2005).
To examine the common ground of the present taxonomic work
with established models of personality pathology in adulthood, we
will first review both categorical and dimensional perspectives on
adult personality disorders.
Categorical and Dimensional Perspectives on Personality
Disorders in Adulthood
Recently, Trull and Durrett (2005) reviewed categorical and
dimensional approaches of personality pathology, discussing their
advantages and problematic features. The DSM–IV (American
Psychiatric Association, 1994) currently distinguishes 10 discrete
personality disorders on Axis II, providing descriptions of charac-
teristic disorder symptoms in adulthood. Although a categorization
of personality disorders facilitates communication among practi-
tioners and further helps to establish guidelines regarding appro-
priate treatment, the accuracy and validity of Axis II diagnoses
have been seriously questioned. Major problematic issues are (a)
the comorbidity among personality disorder diagnoses (Livesley,
2003; Skodol, Rosnick, Kellmann, Oldham, & Hyler, 1988); (b)
the boundary of Axis I and Axis II (Clark, Vorhies, & McEwen,
2002; Widiger, 2003), reflected in the extensive co-occurrence of
Barbara De Clercq, Filip De Fruyt, Karla van Leeuwen, and Ivan
Mervielde, Department of Developmental, Personality and Social Psychol-
ogy, Ghent University, Gent, Belgium.
This research was supported by Special Research Fund of Ghent Uni-
versity Grant 011D0201 awarded to Barbara De Clercq under supervision
of Filip De Fruyt and Ivan Mervielde. We thank Leen De Medts for
collecting data.
Correspondence concerning this article should be addressed to Barbara
De Clercq, Department of Developmental, Personality and Social Psychol-
ogy, Ghent University, H. Dunantlaan 2, B-9000, Gent, Belgium. E-mail:
barbaraj.declercq@ugent.be
Journal of Abnormal Psychology Copyright 2006 by the American Psychological Association
2006, Vol. 115, No. 4, 639 – 657 0021-843X/06/$12.00 DOI: 10.1037/0021-843X.115.4.639
639
personality disorders and other mental disorders; and (c) the lack
of a meaningful point of demarcation differentiating normal from
abnormal psychological functioning (Widiger & Clark, 2000). The
overlap of Axis I and Axis II creates difficulties in discriminating
trait pathology from other types of psychopathology and addition-
ally suggests that both axes have a common etiology (Klein &
Schwartz, 2002; Tyrer, Gunderson, Lyons, & Tohen, 1997).
The dimensional point of view describes both Diagnostic and
Statistical Manual of Mental Disorders (American Psychiatric
Association, 1980, 1987, 1994) axes along continua reflecting
similar underlying mechanisms that contribute to temporary, epi-
sodic forms of psychopathology (Axis I) as well as to more
enduring behavior patterns of Axis II disorders (Krueger & Tack-
ett, 2003; Tillfors, Furmark, Ekselius, & Fredrikson, 2004; Tyrer et
al., 1997). Dimensional advocates further assume that diagnostic
overlap within Axis II partly results from common underlying
traits affecting different personality diagnoses (Livesley, Schroe-
der, Jackson, & Jang, 1994). They also consistently argue that
personality disorders do not reflect discrete categories but continua
of functioning, where the boundaries between adaptive and mal-
adaptive functioning are not clear (Saulsman & Page, 2004). Traits
are described as varying from normal to extreme, and personality
disorders are conceived as extreme variants of normal personality
functioning. The dimensional viewpoint has been empirically sup-
ported throughout the past years (Widiger & Frances, 2002),
suggesting that a confined set of dimensions accounts for the
common features of Axis II personality disorders (Trull & Durrett,
2005). A variety of dimensional models has been proposed for the
description of personality pathology in adulthood from both adap-
tive and maladaptive trait perspectives.
Considering a maladaptive trait perspective, two research teams
independently adopted different bottom-up strategies to define the
personality pathology symptom domain and examined the under-
lying structure of personality pathology. Clark (1990) asked clini-
cians to sort Diagnostic and Statistical Manual of Mental Disor-
ders—Third Edition and non-Diagnostic and Statistical Manual of
Mental Disorders—Third Edition (American Psychiatric Associa-
tion, 1980) symptom criteria of personality disorders and criteria
for selected Axis I disorders with a traitlike component into
synonym groups. Alternatively, Livesley (1986) used clinicians’
judgments of the most prototypic features of each personality
disorder based on a list of trait descriptors and behavioral acts
derived from an extensive review of the personality pathology
literature. Both approaches resulted in hierarchically organized
dimensional systems. Clark’s (1993) Schedule for Nonadaptive
and Adaptive Personality (SNAP) distinguishes three higher order
dimensions—that is, Positive and Negative Affect and Disinhibi-
tion—including 22 lower level scales. Livesley’s (1990) Dimen-
sional Assessment of Personality Pathology—Basic Questionnaire
(DAPP–BQ) differentiates four major dimensions—that is, Emo-
tional Instability, Dissocial Behavior, Inhibitedness, and Compul-
sivity—and 18 more specific personality pathology scales. The
hierarchical structure of both models describes personality pathol-
ogy at a broad and specific level, emphasizing the similarities
among disorders (covariation among disorders due to the same
etiologic factor) but also the unique characteristics of disorders
(caused by etiologic factors that are specific for personality
disorders).
Personality disorders were also examined from the perspective
of adaptive trait variation. The majority of studies examined as-
sociations between domains and facets of the five-factor model
(FFM; Costa & McCrae, 1992) and Axis II categories (Costa &
Widiger, 1994; Widiger & Costa, 2002). Consistent links between
FFM traits and personality disorders in both referred and nonre-
ferred samples were observed, even when relying on clinician’s
descriptions of personality disorders (Samuel & Widiger, 2004). In
the search for consensus on the dimensional structure of person-
ality disorders and the association with normal personality char-
acteristics, O’Connor (2005) suggested that a four-component
solution is the most appropriate representation of the underlying
structure, with the four components being congruent with dimen-
sions of the FFM. A recent meta-analysis by Saulsman and Page
(2004) showed that Neuroticism and Agreeableness are the most
prominent FFM factors explaining personality disorder variance.
Although these associations at the domain level are important to
account for the common core across disorders (Morey et al., 2002),
different authors (De Fruyt, De Clercq, Van de Wiele, & Van
Heeringen, 2006; Dyce & O’Connor, 1998; Saulsman & Page,
2004; Trull, Widiger, & Burr, 2001; Widiger, Trull, Clarkin,
Sanderson, & Costa, 2002) argued that FFM facets better differ-
entiate among personality disorders.
Studies relating adaptive traits to the SNAP and DAPP–BQ
dimensions show that personality pathology is partly explained by
the FFM. The SNAP higher order dimensions are significantly
related to Neuroticism, Extraversion, and Conscientiousness
(Clark, Vorhies, & McEwen, 1994), and the DAPP–BQ higher
order factors are associated with the FFM domains Neuroticism,
Agreeableness, Extraversion, and Conscientiousness (Livesley,
Jang, & Vernon, 1998; Schroeder, Wormworth, & Livesley, 1992).
The joint higher order structure of the DAPP–BQ and the SNAP
scales evinces conceptual and empirical relations with four of the
FFM dimensions, that is, Neuroticism, Agreeableness, Extraver-
sion, and Conscientiousness (Clark, Livesley, Schroeder, & Irish,
1996). Markon, Krueger, and Watson (2005) confirmed that nor-
mal and abnormal personality can be described within the same
structural model, with the FFM as the basic framework for de-
scribing normal and abnormal trait variation.
Childhood Personality and Psychopathology
Although personality differences were intensively studied in
past decades, neither the categorical nor the dimensional approach
has specifically addressed childhood antecedents of maladaptive
personality functioning (Widiger & Clark, 2000). Personality psy-
chologists mainly focused on adaptive child personality and the
relation with adult personality functioning, underscoring the prom-
inent role of early trait differences for adult personality develop-
ment (Caspi, Roberts, & Shiner, 2005; Shiner, 2000; Shiner &
Caspi, 2003; Shiner, Masten, & Roberts, 2003). The FFM is
considered to be a valid and comprehensive taxonomy for describ-
ing personality differences in childhood and adolescence (Buyst,
De Fruyt, & Mervielde, 1994; Digman, 1989; Digman & Inouye,
1986; John, Caspi, Robins, Moffitt, & Stouthamer-Loeber, 1994;
Kohnstamm, Halverson, Mervielde, & Havill, 1998; Lamb,
Chuang, Wessels, Broberg, & Hwang, 2002), and significant as-
sociations have been reported between FFM traits in childhood and
640
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
adult personality and adaptation (Caspi, 2000; Shiner & Masten,
2002; Shiner et al., 2003).
The relationships between specific lower level FFM traits and
DSM–IV personality disorder symptoms were recently examined
in two community samples of adolescents (De Clercq & De Fruyt,
2003; De Clercq, De Fruyt, & Van Leeuwen, 2004). The previ-
ously observed associations between adaptive FFM facets and
categorical Axis II disorders in adulthood (Trull et al., 2001) were
largely replicated in adolescence, suggesting a continuous pattern
of adaptive–maladaptive trait covariation in these life stages (Mer-
vielde, De Clercq, De Fruyt, & Van Leeuwen, 2005). No studies,
however, addressed the adaptive–maladaptive trait associations in
childhood, because DSM–IV does not diagnose personality disor-
ders in children and only identifies childhood psychopathology on
Axis I (American Psychiatric Association, 1994).
Researchers focusing on child psychopathology or Axis I dis-
orders demonstrated that problems in childhood often continue in
adulthood (Briggs-Gowan et al., 2003; Cohen et al., 1993; Good-
win et al., 2004; Pine et al., 1999; Poulton et al., 2000; Roza et al.,
2003). Costello, Egger, and Angold (2005) concluded in a recent
review that many lifetime psychiatric disorders indeed first appear
in childhood or adolescence. Several researchers also underscored
the importance of childhood Axis I disorders for the development
of Axis II symptoms in adulthood (Lewinsohn, Rohde, Seeley, &
Klein, 1997). Kasen and colleagues (1999) suggested that child-
hood psychiatric disorders may predispose an individual to a more
persistent form of Axis II pathology in adulthood. Ramklint, von
Knorring, von Knorring, and Ekselius (2003) showed that major
depressive, disruptive, and substance-related disorders during
childhood and adolescence are significant predictors of adult per-
sonality disorders. These findings all provide compelling evidence
for the continuity of psychopathology from early life, eventually
indicating that underlying trait characteristics play an important
role in processes of maladjustment over time.
The availability of a comprehensive taxonomy that specifically
describes maladaptive traits in childhood can clarify the long-term
importance of early personality symptoms to explain maladaptive
patterns of personality in adulthood. It further creates opportunities
to directly examine the relation between adaptive and maladaptive
traits in childhood and to study child psychopathology from an
age-specific maladaptive trait perspective.
Westen, Shedler, Durrett, Glass, and Martens (2003) were
among the first to construct a specific and empirically grounded
classification of personality pathology for adolescents (i.e.,
Shedler–Westen Assessment Procedure-200; SWAP-200-A). This
SWAP-200-A is a Q-sort instrument for adolescents that was
developed by deleting, revising, and adding items of the adult
Shedler–Westen Assessment Procedure (Westen & Shedler,
1999a, 1999b) version, using various techniques and relying on
different sources of information. Although the results support the
validity of the SWAP-200-A as a valid assessment aid for adoles-
cent personality pathology, the construction heavily relies on
adapting adult personality disorder symptom items for younger age
groups and hence represents a top-down approach to the construc-
tion of personality descriptive models.
The top-down approach may not be the most optimal to study
developmental antecedents of personality disorders and their links
with adaptive traits, given that phenotypic expressions of person-
ality pathology can be expected to be different across the life span,
especially for younger ages (De Clercq et al., 2004). A bottom-up
strategy (De Fruyt, Mervielde, Hoekstra, & Rolland, 2000), start-
ing from the full range of trait-related symptoms manifested in
childhood, may suggest a different set of dimensions representing
trait pathology at younger ages, compared with the models of
Clark (1993) and Livesley (1990) for adults. The first objective of
Study 1 was therefore the compilation and structuring of an age-
specific item pool (Dimensional Personality Symptom Item Pool;
DIPSI; De Clercq, De Fruyt, & Mervielde, 2003) of trait-related
symptoms in childhood. Item compilation strategies were theory
driven, whereas structuring of the taxonomy was empirically
based. Parallel to adult maladaptive taxonomies (Clark, 1993;
Livesley, 1990), we hypothesized that childhood manifestations of
personality pathology can be hierarchically organized, with spe-
cific behavioral, emotional, and cognitive characteristics struc-
tured under more general maladaptive personality dimensions. The
second objective was to examine the nature of the structure of the
DIPSI at different levels of the hierarchy. The third objective was
to compare the DIPSI structure with adult personality pathology
models, in order to explore similarities between child and adult
maladaptive trait taxonomies. The final objective of Study 1 was to
investigate the concurrent validity of the DIPSI, describing the
associations with psychopathology and adaptive personality traits.
Study 1: Construction of the DIPSI
Method
Participants and Procedures
Referred children. Two hundred and five boys (n ⫽ 118) and girls
(n ⫽ 87) were recruited by 3rd-year undergraduate psychology students
from Ghent University. Students were requested to solicit a child consult-
ing an outpatient treatment program in general mental health services,
including psychiatric clinics, pediatric units of hospitals, services for
school counseling, private psychotherapy services, and psychomedical
services for children with developmental and learning disorders. Students
contacted a psychologist or psychiatrist from an online list with registered
providers of psychological care all over Flanders, Belgium. Psychologists
and psychiatrists selected one of their clients to participate in the study,
following the chronological order of their appointment schedule. Children
were between 6 and 14 years, and exclusion criteria were the presence of
a physical disability or a condition of chronic disease. Students visited the
families at home and asked mothers to complete a set of questionnaires.
The treating pediatric psychologist provided information regarding the
initial reason for counseling. Students were instructed not to assist partic-
ipants to ensure that their independent opinion was assessed. All mothers
were assured that the information would be treated as confidential and
would only serve research purposes. Written informed consent was ob-
tained from all mothers and psychologists at the moment of assessment.
The children’s mean age was 9.91 years (SD ⫽ 1.91), ranging from 5.5 to
14.5 years. The period of psychological treatment ranged from 0 to 56
months, with a mean duration of 10 months (SD ⫽ 10.13). Children’s
primary reason for referral included a variety of behavioral and emotional
problems, with anxiety and depressive symptoms as the primary reason for
referral in 21.0% of the children; 24.9% presented externalizing problem
behavior like lying, aggression, and temper tantrums; 15.6% of the children
exhibited enduring behavioral and emotional difficulties related to major
stress events (divorce or death of the parents); 6.8% struggled with psy-
chosomatic complaints (pain without physical cause, eating and sleeping
problems); 9.3% suffered from attention and concentration problems with-
out neurological dysfunction; 11.2% demonstrated behavioral problems
641
PERSONALITY PATHOLOGY IN CHILDHOOD
due to developmental disorders such as attention-deficit/hyperactivity dis-
order, Tourette’s syndrome, and autism spectrum disorder; 8.3% of the
children showed withdrawn behavior or defective social skills; 1% exhib-
ited obsessive– compulsive behavior; and 0.5% had symptoms of auto
mutilation and suicidal thoughts. For 1.5% of the children, no information
on the initial reason of counseling was available.
Nonreferred children. First- and 2nd-year undergraduate psychology
students from Ghent University collected, for course requirements, data
from children of the general population. First-year students were given
addresses of families with children attending 1st-year primary school who
already participated in longitudinal research.
1
Mothers completed a number
of inventories, including the DIPSI. Second-year students were asked to
recruit 2 children between 8 and 14 years of the same family, as part of a
sibling–family study. Mothers completed the DIPSI for both siblings.
Students visited all participants at home and gave detailed instructions
about how to complete the inventory. Informed consent was obtained from
all mothers. The sample (N ⫽ 749) included 382 boys and 367 girls, with
a mean age of 7.54 years (SD ⫽ 2.18), ranging from 5.2 to 15 years.
Almost all children attended normal primary or secondary schools, with
only 1% enrolled in schools for children with learning disabilities.
Measures
Compilation of the DIPSI: Theoretical strategies. Two different the-
oretical strategies, a bottom-up and a top-down approach, were used to
ensure a broad and comprehensive coverage of maladaptive traits in
childhood. The bottom-up approach corroborates previous work of Mer-
vielde and De Fruyt (1999, 2002) constructing a trait taxonomy of the
normal range of personality differences for children, supplemented with a
methodology suggested by Trull and Widiger (1997) to alter adaptive trait
items into maladaptive variants. Mervielde and De Fruyt (1999) assembled
a large pool of parental free personality descriptors of children aged 6 to 12
and represented this descriptive content in the Hierarchical Personality
Inventory for Children (HiPIC; Mervielde & De Fruyt, 1999, 2002; Mer-
vielde, De Fruyt, & De Clercq, 2005). The HiPIC assesses five broad
personality domains that closely resemble the dimensions of the FFM in
adults, that is, Extraversion, Benevolence, Conscientiousness, Emotional
Stability, and Imagination. The HiPIC consists of 18 facets: Extraversion
includes subscales to describe Energy, Expressiveness, Optimism, and
Shyness; Benevolence comprises Altruism, Dominance, Egocentrism,
Compliance, and Irritability subscales; Concentration, Perseverance, Or-
derliness, and Achievement motivation are structured under Conscientious-
ness; Emotional Stability includes Anxiety and Self-Confidence subscales;
and finally Imagination comprises Creativity, Intellect, and Curiosity sub-
scales. All facets are represented by eight items, with a similar grammatical
format, formulated in the third-person singular, without negations in the
item, and excluding personality descriptive adjectives. Furthermore, all
items refer to concrete observable behavior. The HiPIC can thus be
considered as a comprehensive measure of trait differences that can be
reliably observed in childhood.
Trull and Widiger (1997) suggested that extreme variants of adaptive
traits are useful for the description of maladaptive functioning, assuming
that the extremes of both poles hinder optimal functioning of the individual
in particular contexts or situations. Widiger and colleagues (2002) provided
descriptions of personality disorder symptoms in terms of combinations of
extreme scores on the NEO Personality Inventory—Revised (Costa &
McCrae, 1992) facets. Haigler and Widiger (2001) found that the explained
personality disorder variance increased when NEO Personality Inven-
tory—Revised facet items were explicitly revised to suggest maladaptive-
ness. The same idea was adopted in the present work, that is, constructing
maladaptive variants of HiPIC items that covered the extremes of adaptive
personality facets and reflected concrete trait-related symptoms in children.
For example, the HiPIC item “has a sense of order” was transformed to
maladaptive descriptors such as “always leaves everything lying around”
(low orderliness) and “is obsessed with clearing everything” (high
orderliness).
Although Trull and Widiger (1997) suggested that maladaptive variants
of all personality domains are significantly related to adult personality
dysfunctioning, it was difficult to write maladaptive descriptors for the
HiPIC Imagination domain, referring to pathological variants of creativity,
intellect, or curiosity in childhood. Given that this fifth factor (a) is not
represented in the temperament literature, (b) explains only minor variance
in adolescent and adult psychopathology (Mervielde et al., 2005), and (c)
is generally absent in adult personality pathology measures (Widiger &
Simonsen, 2005), no maladaptive variants for facets of this personality
domain were included in the item pool.
The pool of maladaptive HiPIC descriptors was extended by a top-down
strategy, screening for additional descriptors in two instruments that assess
personality disorders in adults, that is, the Structured Clinical Interview for
DSM–IV Axis II Personality Disorders (SCID-II; First, Gibbon, Spitzer,
Williams, & Benjamin, 1997) and the Assessment of DSM–IV Personality
Disorders (ADP-IV; Schotte & De Doncker, 1996; Schotte et al., 2004;
Schotte, De Doncker, Vankerckhoven, Vertommen, & Cosyns, 1998). The
SCID-II is an internationally used and recommended interview to assess
the DSM–IV criteria for personality disorders in adulthood, whereas the
ADP-IV is an inventory developed in Flanders describing DSM–IV Axis II
symptoms enabling a within-language comparison with the maladaptive
HiPIC descriptors. On the basis of clinical judgment, Barbara De Clercq,
Filip De Fruyt, and Ivan Mervielde screened these two instruments for item
content that was applicable to describe personality symptoms in childhood.
The bottom-up and top-down approaches resulted in a pool of 506
descriptors, including 333 descriptors derived from HiPIC items
2
and 83
ADP-IV and 90 SCID-II descriptors. A large number of ADP-IV and
SCID-II items was designated as applicable to children in order to ensure
that no relevant content was omitted from the total item pool prior to
conducting structural analyses. The total pool was subsequently classified
in 44 content-based descriptor sets. Thirty-three descriptor sets represented
maladaptive variants of HiPIC facets and SCID-II/ADP-IV equivalents,
whereas the 11 remaining descriptor sets grouped personality disorder
symptoms that only contained items sampled from the SCID-II or the
ADP-IV. Barbara De Clercq initially assigned all 506 descriptors to the 44
descriptor sets, and Filip De Fruyt and Ivan Mervielde subsequently
inspected and refined this classification, until consensus among the three
was achieved. Finally, 4 – 6 items were written for each personality pathol-
ogy descriptor set, resulting in a pool of 256 items.
3
Items were formulated
in the third-person verb form, avoiding negations and using a yes–no
response format.
Structuring of the DIPSI: Empirical procedures. In a first step, reli-
ability analysis was conducted at the level of the 44 item sets, analyzing
maternal reports for the sample of referred children (N ⫽ 205). Items
considerably lowering internal consistency of item sets were reassigned on
the basis of their correlations with all other item sets. Items were only
reassigned when this increased the reliability of the new set. Items that
1
Because of the design, the age range for this sample was restricted to
6 years.
2
For some HiPIC facets, we constructed several specific maladaptive
item sets (e.g., the maladaptive item sets of the HiPIC facet Anxiety
include both general Anxious traits, Separation Anxiety, Depressive traits,
and Rumination). The decision to construct these additional item sets was
based on clinical judgment of Barbara De Clercq, Filip De Fruyt, and Ivan
Mervielde.
3
It may give the impression that a large number of descriptors were
dropped from the item pool. These descriptors were considered not con-
crete enough or their content was already represented in another more
appropriate descriptor.
642
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
could not be reallocated to any other item set (i.e., items with item total
correlations below .30) were dropped.
In order to provide an empirical basis for the item set organization, we
conducted five separate item-level principal-axis factorings with oblique
rotation on subgroups of item sets: Four principal-axis factorings were
conducted on the items enclosed in the sets developed for each HiPIC
domain, and one principal-axis factoring was conducted on the items of the
sets derived from the adult personality disorder criteria (see Table 1 for an
overview of the subgrouping of item sets). Because such analyses require
a larger sample for an optimal subjects:variables ratio, we relied on
maternal DIPSI reports of the nonreferred sample of children (N ⫽ 749).
In order to obtain item sets that measure unique maladaptive trait variance,
we initially omitted items with high cross-loadings (⬍.30
4
) as well as items
with loadings less than .30 on any of the factors from that particular item
set. High-loading items (⬍.30) on a different factor than originally pre-
sumed were only reassigned when the Pearson product–moment correla-
tion coefficient of that item with the new item set was greater than .30 and
only if the item did not lower the alpha reliability coefficient of that item
set. Omitted items were subsequently correlated with all other item sets,
and the same criteria were applied for further reassignment of items. This
procedure for eliminating and reassigning items was repeated until all
items loaded above .30 on a factor representing a specific item set, showed
no cross-loadings on item sets representing different maladaptive trait
components, and contributed to the reliability of the item set.
The unidimensionality of the final item sets was explored in a third step
using item-level principal-axis factoring within each item set. Item sets for
which unidimensionality was demonstrated are from here on referred to as
DIPSI facets.
Different principal-axis factoring of the DIPSI facets were conducted in
order to examine the higher order structure of the facets, and Pearson
product–moment correlation coefficients were calculated to explore rela-
tionships between factors at different levels of the hierarchy (Objective 2).
In order to examine the comprehensiveness of the DIPSI taxonomy (Ob-
jective 3), we compared the DIPSI facets with the structure of the adult
DAPP–BQ (Livesley, 1990) and the SNAP (Clark, 1993), matching the
DIPSI facets with the DAPP–BQ/SNAP scales on the basis of content
similarity (Clark & Livesley, 2002). Finally, Pearson product–moment
correlations between the DIPSI higher order scale scores and the HiPIC
and the Dutch version of the Child Behavior Checklist (CBCL; Verhulst,
Van der Ende, & Koot, 1996) were calculated to explore the associations
with adaptive personality and psychopathology measures (Objective 4).
Instruments. Mothers of children in both samples completed the
DIPSI. Mothers of the referred group additionally provided ratings on the
HiPIC (Mervielde & De Fruyt, 1999) and the CBCL (Verhulst et al., 1996).
4
The value of .30 has been traditionally used as a cutoff in factor
analysis (Lemery, Essex, & Smider, 2002).
Table 1
Overview of Item Sets Derived From the Hierarchical Personality Inventory for Children
(HiPIC), the Structured Clinical Interview for DSM–IV Axis II Disorders (SCID-II), and the
Assessment of DSM–IV Personality Disorders (ADP-IV) After Theoretical Compilation
Procedures
Original instrument DIPSI item set
HiPIC
Emotional Stability
Anxiety Anxious Traits, Separation Anxiety, Depressive Traits,
Rumination
Self-Confidence Lack of Self-Confidence, Dependency, Naivety
Extraversion
Energy Hyperactive Traits, Impatience, Risk Taking, Passivity
Expressiveness Hyperexpressive Traits, Withdrawn Traits
Optimism Apathy
Shyness Shyness, Social Insecurity, Social Avoidance
Benevolence
Altruism Lack of Empathy
Dominance Dominance, Manipulative Traits, Submissiveness
Egocentrism Egocentrism
Compliance Resistance
Irritability Irritability, Aggressive Traits
Conscientiousness
Concentration Distraction
Perseverance Lack of Perseverance
Order Extreme Order, Disorderliness, Perfectionism, Inflexibility,
Obsessive Traits
Achievement motivation Extreme Achievement Striving
SCID-II/ADP-IV
Dependent personality disorder criteria Insecure Attachment
Paranoid personality disorder criteria Paranoid Traits, Unforgivingness
Schizotypal personality disorder criteria Cognitive Distortion
Narcissistic personality disorder criteria Grandiosity, Narcissistic Traits
Antisocial personality disorder criteria Remorselessness
Borderline personality disorder criteria Affective Lability, Impulsivity, Ineffective Coping
Histrionic personality disorder criteria Overreactivity
Note. Italicized Dimensional Personality Symptom Item Pool (DIPSI) facets were dropped after the reassign-
ment process. DSM–IV ⫽ Diagnostic and Statistical Manual of Mental Disorders (4th ed.).
643
PERSONALITY PATHOLOGY IN CHILDHOOD
The CBCL assesses children’s problem behavior on a 3-point Likert scale.
Problem behaviors are scored on syndromes (Withdrawn, Somatic Com-
plaints, Anxious/Depressed, Social Problems, Thought Problems, Atten-
tion Problems, Delinquent Behavior, and Aggressive Behavior) and broad-
band scales, including an Internalizing, Externalizing, and Total Problem
scale.
Results
Internal Consistency
Alpha coefficients for the initial 44 item sets ranged from .49
(Passivity) to .86 (Emotional Lability), with a median value of .78.
Thirteen items were reassigned to other item sets, and 34 items out
of 256 were omitted because they showed low correlations (⬍.30)
with any of the item sets or did not empirically cluster together to
form a new item set. Four item sets—that is, Apathy, Naivety,
Passivity, and Remorselessness—were removed at this stage. This
first empirical step resulted in an item pool of 222 items, organized
in 40 item sets of trait-related pathology. Table 1 presents an
overview of the item sets derived from the HiPIC and the SCID-
II/ADP-IV, respectively, after optimal assignment of items and
with the omitted item sets italicized.
Exploratory Factor Analysis of Subgroups of Items
The item-level principal-axis factoring of the Emotional
Stability-based sets revealed five clearly interpretable factors, with
the items of Anxious Traits and Rumination loading on the same
factor, and the items of Dependency, Separation Anxiety, Depres-
sive Traits, and Withdrawn Traits loading on different factors. Six
items with low loadings were omitted as well as six items with
high cross-loadings. The factor analysis of Extraversion-based
item sets showed again a five-factor solution with the items of
Shyness, Social Insecurity, and Social Avoidance primarily load-
ing on the same factor; items of Impatience and Hyperactive Traits
loading together; and Withdrawn Traits, Risk Taking, and Hyper-
expressivity items each representing a different factor. Eleven
items demonstrated loadings less than .30 and were eliminated,
together with two items that lowered the reliability coefficient of
their item sets. The factor analysis of the items of the Benevolence-
related sets showed also a five-factor solution, with Aggressive
Traits and Irritability loading together; Dominance and Egocen-
trism items loading a factor; and items of Lack of Empathy,
Submissiveness, and Resistance representing different factors. The
original item set of Manipulative Traits loaded partially on the
Dominance–Egocentrism factor and partially on the Resistance
factor. Six items had loadings below .30, and three items lowered
the reliability coefficient of their respective item sets. The factor
analysis of the items of the Conscientiousness-related sets revealed
six factors, with Lack of Perseverance partially loading Disorder-
liness and partially Distraction, with Inflexibility and Obsessive
Traits loading together, and items of Extreme Achievement Striv-
ing, Order, and Perfectionism each loading separate factors. Three
items were eliminated because of low loadings or lower alpha
coefficients when enclosed in their item sets. The factor analysis of
the items of sets representing adult personality disorder criteria—
beyond the maladaptive HiPIC item sets—resulted in a six-factor
solution, with items of Grandiosity and Narcissistic Traits loading
together; Ineffective Stress Coping and Overreactivity items load-
ing a factor; and items from Affective Lability, Paranoid Traits,
Impulsivity, and Insecure Attachment representing different fac-
tors. Thirteen items showed loadings below .30, two items had
high cross-loadings, and two items lowered the reliability coeffi-
cient of their respective item sets. Two items of the original
Unforgivingness item set were assigned to Lack of Empathy, given
their high loadings on this factor; one item of the original Apathy
item set was assigned to Withdrawn Traits; and one item of the
Narcissistic item set was assigned to Hyperexpressive Traits. The
remaining items (n ⫽ 50) were dropped from the item pool. Table
2 presents the DIPSI lower level structure, resulting from the
implementation of these procedures, with 172 items organized in
27 sets showing satisfactory reliability and mean interitem
correlations.
Exploratory Factor Analyses Within Item Sets
Principal-axis factoring with oblique rotation clearly under-
scored the unidimensional structure of 25 out of 27 sets, except for
Aggressive–Irritable Traits and Ineffective Stress Coping–
Overreactivity suggesting two dimensions. Although the eigenval-
ues of these subdimensions were above 1 in both cases, almost all
items of the two dimensions showed substantial cross-loadings.
The dimensions were also highly correlated (⬍.60), and more than
30 iterations were required to obtain a two-dimensional solution. It
was therefore decided not to split these item sets and to consider
them also as unidimensional facets. The last column of Table 2
presents the range of loadings of the items of each facet, indicating
substantial loadings for all facets.
Hierarchical Structure
Principal-axis factoring at the level of the 27 facets, followed by
Varimax rotation, showed a decreasing pattern of eigenvalues
(9.340, 2.230, 1.929, 1.405, 1.238, 1.007, 0.824, 0.788, . . .) that
suggested to retain either a two-, three-, four-, or five-factor
solution, explaining, respectively, 38.53%, 44.18%, 48.01%, and
51.36% of the total variance. Each of these solutions is presented
in Figure 1 and shows that the DIPSI facets can be organized at
different higher order levels, with lower levels merging into higher
order factors, and with only three facets shifting to a different
higher order factor than expected from the lower order grouping of
facets. Such shifts were observed for the Ineffective Coping, Lack
of Empathy, and Inflexibility facets and are marked with dashed
lines.
The five-factor solution partly resembles factor structures of
adult personality pathology inventories (Clark et al., 1996) pre-
senting a Disagreeable factor (13.55%), a Neuroticism factor
(11.53%) mainly described by the categories Dependency and
Submissiveness, an Affective Instability factor (11.46%), an In-
troversion factor (7.47%), and a Compulsivity factor (7.34%). The
four-factor solution resembles the basic dimensions of adult per-
sonality pathology proposed by Livesley et al. (1998), that is,
Disagreeableness (18.93%), Emotional Instability (14.27%), Intro-
version (7.37%), and Compulsivity (7.44%), with Disagreeable-
ness incorporating the Affective Instability factor of the five-factor
solution, except for the Ineffective Coping facet that shifts to the
Neuroticism factor. The three-factor solution resulted in a Dis-
agreeableness factor that is nearly identical to the Disagreeable-
644
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
ness factor of the four-factor structure, with only Inflexibility
shifting to the Negative Emotionality factor (19.10%), a Negative
Emotionality factor (17.63%) that integrates the Neuroticism and
the Introversion factor of the four-factor solution (except for the
Lack of Empathy facet that shifts to Disagreeableness), and a
Compulsivity factor (7.45%). The two-factor solution clearly rep-
resented Externalizing (20.54%) and Internalizing (17.99%) trait
factors, with the Internalizing factor incorporating the Negative
Emotionality and the Compulsivity factor of the three-factor solu-
tion. All loadings of each solution are significant at p ⬍ .01 (for
N ⫽ 749, loading of ⱖ 兩.18兩 is significant at ␣⫽.01; Stevens,
1996), and almost all loadings of each solution share enough
Table 2
Dimensional Personality Symptom Item Pool (DIPSI) Facets After Theoretical and Empirical Compilation Procedures
DIPSI facet Sample items N
items
M interitem
correlation
coefficient ␣ Range
loadings
Hyperexpressive Traits Talks all the time about his/her own experiences 8 .34 .79 .41–.65
Exhibits his/her inner feelings at all occasions
Hyperactive Traits Can never sit still 7 .45 .85 .51–.83
Gets very annoyed when he/she needs to wait somewhere
Dominance–Egocentrism Always demands other children to listen 8 .31 .77 .49–.65
Manipulates other children repeatedly to have his/her way
Impulsivity Acts constantly without thinking about the consequences 4 .54 .82 .66–.81
Always makes decisions in a very inconsiderate way
Irritable–Aggressive Traits Gets frequently out of control when he/she is angry 9 .37 .83 .52–.71
Gets easily irritated
Disorderliness Always makes a big mess of everything 8 .38 .83 .53–.70
Never takes care of his/her belongings
Distraction Never finishes his/her work 7 .37 .80 .47–.83
Can only be focused for a few moments
Risk Taking Is very attracted to dangerous situations 6 .38 .78 .45–.75
Likes to take risks
Narcissistic Traits Fantasizes all the time about being admired by others 8 .29 .75 .48–.63
Considers him/herself more worthy than others
Affective Lability Has frequent changes in mood from one extreme to the other 6 .45 .83 .47–.83
His/her feelings towards others are very changeable
Resistance Disobeys rules all the time 5 .28 .64 .34–.74
Always refuses to do what is asked
Lack of Empathy Feels no emotions when other children get hurt 10 .23 .74 .39–.73
Is never interested in problems of other children
Dependency Needs someone around all the time 5 .35 .72 .51–.64
Can never undertake something without help
Anxious Traits Worries all the time 7 .32 .74 .43–.66
Panics very easily
Lack of Self-Confidence Always has doubts about him/herself 4 .46 .75 .59–.76
Feels less worthy than other children
Insecure Attachment Wants to have his/her parents always around 4 .40 .72 .50–.84
Clings to other people
Submissiveness Obeys other children all the time 8 .30 .75 .44–.73
Always submits to other children
Ineffective Coping Is very sensitive to stress 8 .37 .82 .51–.72
Is easily overwhelmed by his/her emotions
Separation Anxiety Fears constantly to be on his/her own one day 3 .50 .75 .61–.78
Often fears of being abandoned one day
Depressive Traits Feels often empty inside 4 .29 .58 .37–.82
Regrets too often things that happened in the past
Inflexibility Cannot adjust to sudden changes in plans 9 .26 .76 .40–.65
Feels frequently forced to repeat behavioral acts in a certain order
Shyness Fears every contact with other children 8 .32 .76 .46–.78
Always feels uncomfortable when other children are around
Paranoid Traits Is very suspicious towards other children 5 .27 .62 .45–.64
Thinks that other children want to cheat all the time
Withdrawn Traits Always hides his/her feelings 6 .30 .72 .33–.77
Cannot express feelings of affection
Perfectionism Loses a lot of time by doing things too perfectly 5 .24 .61 .42–.65
Wants life to be perfectly organized
Extreme Achievement Striving Wants to shine at everything 4 .47 .77 .64–.76
Always demands him/herself to be the best
Extreme Order Is obsessed by cleaning 6 .29 .68 .46–.66
Feels having control by being orderly all the time
645
PERSONALITY PATHOLOGY IN CHILDHOOD
variance (loading ⬎ .40; Stevens, 1996) with their higher order
factor to be used for the interpretation of that factor. We decided
to further focus on the four- and two-factor structure because these
solutions appear to closely resemble the four-factor structure of
adult personality pathology (Clark & Livesley, 2002) and the
two-factor structure of both adult (Krueger & Piasecki, 2002) and
child (Achenbach, 1991) psychopathology. Table 3 presents the
factor loadings for both solutions, with the primary loadings
marked in boldface.
Content Validity
Inspection of Table 4 shows that most DIPSI facets have clear
conceptual counterparts within the DAPP–BQ and SNAP taxono-
mies. The DIPSI has parallel facets for all DAPP–BQ and SNAP
scales, except for a particular set of DAPP–BQ or SNAP scales
that are rather uncommon in childhood. These are the DAPP–BQ
Self-Harm scale and its SNAP equivalent Suicide Proneness, the
DAPP–BQ component dimensions Inhibited Sexuality and Addic-
tive Behaviors, and criteria of the SNAP symptom clusters De-
pendency and Antisocial Behavior that refer to partner relation-
ships or financial obligations, respectively. Because descriptors
that are not primarily trait related were excluded, the DIPSI does
not provide parallel facets for the DAPP–BQ Depersonalization
and the SNAP Depersonalization–Derealization scales nor for the
Identity Problem and Identity Disturbance scales. The DAPP–BQ
subscales Sadism and Passivity were initially represented within
the DIPSI through the item sets Remorselessness and Passivity but
were finally removed as distinct facets in the item-reassignment
process.
The DIPSI Hyperactive Traits facet has an equivalent in the
SNAP but not in the DAPP taxonomy, whereas the content of two
additional DIPSI facets may only be partly associated with any of
the adult measures. Hyperexpressive Traits may not be fully rep-
resented by the SNAP scale Dramatic Exhibitionism or the DAPP
scale Narcissism, because its content is somewhat broader than
attention-seeking behavior or the need for adulation. This facet can
be located at the low end of Agreeableness and represents traits
like irritable talkativeness and persistently drawing negative atten-
tion from others. The DIPSI facet Inflexibility includes the inca-
pability to adjust to environmental changes, the feeling of being
forced to repeat certain behavioral acts, and the presence of a rigid
behavioral style in daily life and may hence not be fully repre-
sented in the DAPP–BQ component Rigid Cognitive Style.
Concurrent Validity
Maladaptive personality and psychopathology. Table 5 pre-
sents the correlations of the DIPSI scale scores for both the four-
and the two-factor solutions with the CBCL scales. Considering
the four-factor structure, Disagreeableness is primarily positively
related to the CBCL syndrome scales Aggressive Behavior, De-
linquent Behavior, and Attention Problems and also moderately to
Social Problems. Emotional Instability shows strong positive cor-
relations with Anxious–Depressed, Thought Problems, Withdrawn
Behavior, and Social problems. Introversion is strongly related to
the subscale Withdrawn Behavior and Anxious–Depressed and
moderately to Social Problems. Finally, Compulsivity is moder-
ately correlated with Anxious–Depressed. All of the DIPSI factors
demonstrate a positive correlation with the Total Problem score;
Figure 1. Correlations among hierarchical levels of the Dimensional Personality Symptom Item Pool item
pool, based on the sample of nonreferred children (N ⫽ 749). Categories that shift to another factor are marked
in dashed lines. inflex ⫽ Inflexibility; empa ⫽ Lack of Empathy; coping ⫽ Ineffective Coping.
646
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
Emotional Instability, Introversion, and Compulsivity correlate
primarily with the CBCL Internalizing Problem score, whereas
Disagreeableness is primarily associated with the CBCL External-
izing Problem score. For the two-factor solution, the results dem-
onstrate that the DIPSI Internalizing trait factor is especially highly
related to Anxious–Depressed, Withdrawn Behavior, Thought
Problems, and Social Problems. The Externalizing trait factor
shows an identical correlation pattern with Disagreeableness of the
four-factor structure, because the Disagreeableness factor is iden-
tical in the four- and the two-factor solutions.
Maladaptive personality and adaptive personality. An inspec-
tion of the DIPSI correlations with the HiPIC domains (see Table
5) shows that Disagreeableness is mainly negatively related to the
HiPIC domains Benevolence and Conscientiousness, whereas
Emotional Instability correlates negatively with Emotional Stabil-
ity. Introversion is negatively related to Extraversion and Emo-
tional Stability, and Compulsivity shows a significant positive
correlation with Conscientiousness and a negative correlation with
Emotional Stability. For the two-factor solution, the DIPSI Exter-
nalizing trait factor is primarily negatively related to Benevolence
and Conscientiousness and moderately positively related to Extra-
version, whereas the DIPSI Internalizing trait factor shows a strong
negative relation with Emotional Stability and a moderate negative
correlation with Extraversion.
Study 2: Construction of a Model for Personality
Symptoms in Childhood
The previous study suggests that the four-factor hierarchical
structure of the DIPSI is very similar to the higher order dimen-
sions of personality pathology in adulthood (Clark & Livesley,
2002), in addition to conceptual similarities at the lower order
level. The results further demonstrate that a two-superfactor struc-
ture is empirically related to the basic CBCL dimensions of child
psychopathology and conceptually to the basic dimensions of adult
psychopathology (Krueger & Piasecki, 2002). Study 2 intended to
examine the fit of these two–four higher order structures. The first
objective considered the four-factor structure as a basis for the
construction of a model and evaluated the fit of a two–four-factor
structure in the sample of nonreferred children of Study 1 (N ⫽
Table 3
Four- Versus Two-Factor Structure of the Dimensional Personality Symptom Item Pool (DIPSI)
Facets in Nonreferred Children (N ⫽ 749)
DIPSI facet
Four-factor structure
Two-factor
structure
DIS INS COM ITR INT EXT
Hyperexpressive Traits .74 .17 .30 .12 .24 .77
Hyperactive Traits .71 .30 .06 .07 .25 .72
Dominance–Egocentrism .69 ⫺.01 .39 .15 .14 .71
Impulsivity .67 .30 ⫺.08 .07 .22 .67
Irritable–Aggressive Traits .61 .31 .17 .26 .39 .64
Disorderliness .60 .28 ⫺.15 .10 .20 .58
Distraction .58 .44 ⫺.19 .08 .31 .56
Risk Behavior .58 .19 .07 .08 .17 .59
Narcissistic Traits .54 ⫺.06 .42 .06 .09 .55
Inflexibility
a
.51 .47 .25 .33 .59 .55
Affective Lability .50 .32 .17 .23 .14 .54
Resistance .48 .04 .05 .29 ⫺.00 .49
Dependency .23 .64 .06 .12 .60 .25
Anxious Traits .16 .63 .27 .17 .68 .19
Lack of Self-Confidence .11 .60 .10 .17 .61 .14
Insecure Attachment .24 .58 .14 .21 .62 .26
Submissiveness .29 .54 .03 .19 .53 .31
Ineffective Coping .43 .51 .27 .11 .53 .46
Separation Anxiety .11 .50 .10 .09 .49 .12
Depressive Traits .17 .35 .09 .26 .43 .20
Perfectionism .17 .28 .72 .03 .42 .24
Extreme Achievement Striving .12 .09 .66 .09 .29 .19
Extreme Order ⫺.11 .35 .47 .06 .45 ⫺.04
Shyness .14 .19 .06 .80 .47 .22
Paranoid Traits .11 .30 .04 .68 .51 .18
Lack of Empathy
b
.35 .22 .09 .36 .27 .38
Withdrawn Traits .22 .32 .08 .36 .44 .26
Note. Primary loadings are marked in boldface. DIS ⫽ Disagreeableness; INS ⫽ Emotional Instability;
COM ⫽ Compulsivity; ITR ⫽ Introversion; INT ⫽ Internalizing Traits; EXT ⫽ Externalizing Traits.
a
Because Inflexibility shifts from Disagreeableness in the four-factor structure to the Internalizing trait factor in
the two-factor solution, and because of the substantial cross-loading on Emotional Instability, this facet will be
assigned to Emotional Instability for further analyses.
b
Because Lack of Empathy shifts from Introversion in
the four-factor structure to the Externalizing trait factor in the two-factor solution, and because of the substantial
cross-loading on Disagreeableness, this facet will be assigned to Disagreeableness for further analyses.
647
PERSONALITY PATHOLOGY IN CHILDHOOD
749). A second objective focused only on the broader four-factor
model and evaluated its fit in a new independent sample of
nonreferred children (N ⫽ 242) and in the referred sample of Study
1(N ⫽ 205). Given the higher order similarity with adult mal-
adaptive taxonomies, the third objective examined whether the
higher order structure of trait pathology can be extended from
childhood to adolescence, using self-ratings of nonreferred ado-
lescents (N ⫽ 453). A last objective investigated factorial invari-
ance across these groups of referred and nonreferred children and
nonreferred adolescents.
Method
Participants and Procedures
Nonreferred children. Third-year undergraduate psychology students
from Ghent University were invited to administer the DIPSI to mothers of
children of the general population. All students visited participants at home
and gave detailed instructions on how to complete the questionnaire.
Informed consent was obtained from all children and their mothers. The
sample (N ⫽ 242) consisted of 112 boys and 130 girls with a mean age of
10.62 years (SD ⫽ 1.07), ranging from 7.1 to 14 years. Almost all (98.7%)
were enrolled in normal primary or high schools, whereas 0.08% attended
specific educational programs for children with learning disabilities.
Nonreferred adolescents. Third-year undergraduate psychology stu-
dents from Ghent University were required to collect DIPSI self-ratings
from 2 adolescents of the general population, preferably a boy and a girl
between 12 and 15 years old. All students visited participants at home and
gave detailed instructions about how to complete the questionnaire. In-
formed consent was obtained from all adolescents and their parents. The
sample (N ⫽ 453) consisted of 216 boys and 237 girls, with a mean age of
13.9 years (SD ⫽ 1.14), ranging from 10.5 to 16.7 years. Almost all
(98.6%) were enrolled in normal primary or high schools, whereas 1.4%
attended specific educational programs for children with learning
disabilities.
Model Construction: Data-Analytic Strategy
Gerbing and Hamilton (1996) and Widaman (1993) suggested using
exploratory factor analysis to explore the form of a multiple-indicator
model followed by confirmatory factor analysis to evaluate model fit and
provide parameter estimation. They demonstrated that principal-axis fac-
toring with Varimax rotation
5
is particularly useful for the construction of
multiple-indicator measurement models as a precursor of confirmatory
5
Although the latent factors obtained from confirmatory factor analysis
generally include correlated factors, using oblique rotation would result in
large factor correlations, which often leads toward larger errors of estima-
tion for the indicators of each factor.
Table 4
Comparison of the Dimensional Personality Symptom Item Pool (DIPSI) Lower Level Facets With the Dimensional Assessment of
Personality Pathology—Basic Questionnaire (DAPP–BQ) and the Schedule for Nonadaptive and Adaptive Personality (SNAP) Scales
DIPSI facet DAPP–BQ SNAP
Disagreeableness
Hyperexpressive Traits Narcissism Dramatic Exhibitionism
Hyperactive Traits High Energy
Dominance–Egocentrism Rejection, Interpersonal Disesteem Self-Centered Exploitation
Impulsivity Stimulus Seeking Impulsivity
Irritable–Aggressive Traits Affective Lability Anger–Aggression
Disorderliness Passive Oppositionality Passive Aggressiveness
Distraction Passive Oppositionality Passive Aggressiveness
Risk Taking Stimulus Seeking Impulsivity, High Energy
Narcissistic Traits Narcissism Dramatic Exhibitionism, Grandiose Egocentrism
Affective Lability Affective Lability Instability
Resistance Conduct Problems Antisocial Behavior
Lack of Empathy Interpersonal Disesteem Self-Centered Exploitation
Emotional Instability
Dependency Diffidence Dependency
Anxious Traits Anxiousness Negative Affect, Pessimism
Lack of Self-Confidence Diffidence, Anxiousness Self-Derogation
Insecure Attachment Insecure Attachment Dependency
Submissiveness Diffidence Dependency
Ineffective Coping Brief Stress Psychosis, Affective Lability Hypersensitivity
Separation Anxiety Insecure Attachment Dependency
Depressive Traits Anhedonia Anhedonia, Negative Affect
Inflexibility Rigid Cognitive Style
Introversion
Shyness Social Avoidance Social Isolation
Paranoid Traits Suspiciousness Suspiciousness
Withdrawn Traits Restricted Expression Emotional Coldness
Compulsivity
Perfectionism Compulsivity Conventionality–Rigidity
Extreme Achievement Striving Compulsivity Conventionality–Rigidity
Extreme Order Compulsivity Conventionality–Rigidity
Note. Only the DAPP–BQ/SNAP scales that have a DIPSI equivalent are reported.
648
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
factor analysis. Following this recommendation, we relied on the Varimax
rotated factor solution presented in Table 3 to construct the model. Con-
sidering that at least three indicators are required for the identification of a
latent factor (Jo¨reskog & So¨rbom, 1989), we selected three facets for each
higher order factor in two steps: In order to represent the broad content of
each higher order factor, we first conducted principal-axis factoring with
oblimin rotation within each factor and screened for interpretable subdo-
mains each to be represented by a specific marker. Second, each of the
factor scores was regressed stepwise on its facets, and the three facets with
the largest standardized beta coefficients were retained as markers.
The ability of this marker model to fit the data of different samples was
subsequently examined using confirmatory factor analysis in Lisrel 8.72
(Jo¨reskog & So¨rbom, 2005). An additional and important issue in the
development of models of personality structure is whether the same struc-
ture can be applied to two or more groups (Finch & West, 1997). To
evaluate this factorial invariance for maternal ratings of nonreferred (N ⫽
242) versus referred (N ⫽ 205) children, and for nonreferred children (N ⫽
242) versus nonreferred adolescents (N ⫽ 453) using maternal ratings and
self-ratings, respectively, we conducted multisample analyses following
the procedure for comparing factor structures outlined by Finch and West
(1997). Because of the nonnormality of the data, all analyses were run on
the asymptotic covariance matrices.
Results
Marker Selection
Exploratory factor analyses of the facets of Emotional Instabil-
ity, Introversion, and Compulsivity suggested only one factor for
each analysis with no interpretable subfactors, indicating that the
facets assigned to these factors represent similar higher order
content characteristics. Stepwise regression analysis for Emotional
Instability showed that Dependency, Insecure Attachment, and
Anxious Traits explain the largest amount of variance. For Intro-
version, the largest beta coefficients were observed for the facets
Shyness, Paranoid Traits, and Withdrawn Traits. Compulsivity is
defined by only three facets that were all selected as markers for
this higher order domain. For Disagreeableness, exploratory factor
analysis revealed three clearly interpretable subdomains, indicat-
ing that the content of this factor is broader than more narrowly
defined factors such as Compulsivity. It was therefore decided to
retain the facet with the highest beta coefficient for each of the
three subdomains, bringing the number of markers for Disagree-
ableness also to three. The first subdomain represents an Affective
Instability factor, with Irritable–Aggressive Traits as the highest
loading facet. The second subdomain corresponds clearly to a
Dominance factor, with Dominance–Egocentrism as the highest
loading facet, and finally the third subdomain describes a
Hyperactivity–Impatience factor, with the DIPSI facet Hyperactive
Traits as the highest loading category. Correlation coefficients of
the DIPSI marker scores and the DIPSI full-scale scores for
nonreferred children (N ⫽ 242) range between .94 (Disagreeable-
ness) and 1.00 (Compulsivity), indicating that the marker structure
closely represents the content of the full DIPSI.
Model Specification
We assumed that 12 measured indicators represent four latent
factors in all samples. Irritable–Aggressive Traits, Dominance–
Egocentrism, and Hyperactive Traits were hypothesized to mark
the Disagreeableness factor; Shyness, Paranoid Traits, and With-
drawn Traits were retained as indicators of the Introversion factor;
Dependency, Insecure Attachment, and Anxious Traits mark the
Table 5
Correlations of the Dimensional Personality Symptom Item Pool (DIPSI) With the Child
Behavior Checklist (CBCL) and the Hierarchical Personality Inventory for Children (HiPIC) in
Referred Children (N ⫽ 205)
Scale
DIPSI four-factor structure
DIPSI two-factor
structure
DIS INS ITR COM EXT INT
CBCL syndromes and dimensions
Withdrawn Behavior .24** .48** .62** .20* .24** .53**
Somatic Complaints .06 .34** .17 .13 .06 .31**
Anxious–Depressed .23* .71** .56** .38** .23* .73**
Social Problems .36** .48** .42** .23* .36** .50**
Thought Problems .25** .50** .33** .22* .25** .49**
Attention Problems .63** .42** .35** .12 .63** .41**
Delinquent Behavior .59** .15 .20* .13 .59** .18*
Aggressive Behavior .77** .24* .20* .15 .77** .25**
Internalizing Problems .22* .67** .57** .32** .22* .68**
Externalizing Problems .77** .23* .21* .16 .77** .25**
Total Problem score .62** .59** .48** .30** .62** .61**
HiPIC domains
Extraversion .24** ⫺.15 ⫺.46** ⫺.01 .24* ⫺.21*
Benevolence ⫺.66** ⫺.04 ⫺.03 .02 ⫺.66** ⫺.04
Conscientiousness ⫺.47** .03 ⫺.02 .45** ⫺.47** .09
Emotional Stability .01 ⫺.64** ⫺.37** ⫺.30** ⫺.01 ⫺.62**
Imagination ⫺.16 ⫺.17 ⫺.14 .11 ⫺.16 ⫺.13
Note. DIS ⫽ Disagreeableness; INS ⫽ Emotional Instability; ITR ⫽ Introversion; COM ⫽ Compulsivity;
EXT ⫽ Externalizing Traits; INT ⫽ Internalizing Traits.
* p ⬍ .01. ** p ⬍ .001.
649
PERSONALITY PATHOLOGY IN CHILDHOOD
Emotional Instability factor; and Extreme Achievement Striving,
Perfectionism, and Extreme Order define the Compulsivity factor.
It was further assumed that Introversion, Emotional Instability, and
Compulsivity represent the superorder factor Internalizing Traits,
whereas Disagreeableness represents the Externalizing trait factor
(see Figure 1). It was presumed that the DIPSI facet markers would
load only their intended factor with zero loadings on the nonin-
tended factors. The measurement error variances of the 12 ob-
served variables were hypothesized to be uncorrelated, and latent
factors were allowed to correlate.
Model Fit
In line with the first objective, we examined whether a four-
factor model can be hierarchically organized in two superfactors in
the sample of nonreferred children (Study 1, N ⫽ 749). A good fit
to the data for the two–four model was found. The absolute fit is
represented by the ratio of the chi-square and its degrees of
freedom, with ratios of less than five used as standard for adequate
fit (Bollen, 1989). The present ratio shows a value of 3.68, indi-
cating an acceptable fit. However, this index has been the subject
of criticism (Finch & West, 1997; Kelloway, 1996; Schermelleh-
Engel, Mossbrugger, & Mu¨ller, 2003) and has been recommended
as a descriptive index of fit rather than as a statistical test (Jo¨res-
kog, 1993). An alternative index of overall fit is the root-mean-
square error of approximation (RMSEA), with a value of .06 in the
present study. The RMSEA was developed by Steiger (1990), who
suggested that values below .10 indicate acceptable fit to the data
(see also Browne & Cudeck, 1993). The comparative fit is repre-
sented by the nonnormed fit index (NNFI; Bentler & Bonnett,
1980) and the comparative fit index (CFI; Bentler, 1990), two
indices that are commonly used in studies with structural equation
analyses (McDonald & Ringo Ho, 2002). Kelloway (1996) sug-
gested that values of the NNFI and the CFI exceeding .90 indicate
an adequate fit to the data. The present results show values of .97
for both the NNFI and the CFI, indicating a good fit. Completely
standardized factor loadings are all significant at p ⬍ .01 and range
from .68 to .78 for Disagreeableness, from .49 to .83 for Introver-
sion, from .59 to .80 for Emotional Instability, and from .55 to .93
for Compulsivity. Completely standardized factor loadings (sig-
nificant at p ⬍ .01) for the two-factor structure range from .49 to
.79 for Internalizing Traits. The Externalizing trait factor is iden-
tical to the Disagreeableness factor; hence, no standardized factor
loadings are reported. Factor correlations show moderate correla-
tions between the subordinate factors loading on the superfactor
Internalizing Traits (with rs ranging from .31 to .63), and the two
superfactors are highly correlated (r ⫽ .79).
For the second objective, the fit of the broader four-factor model
in nonreferred and referred samples of children and adolescents
was examined. Absolute and comparative fit indices are reported
in Table 6 and suggest an acceptable fit of the model to the data for
each of the samples. The
2
/df ratios are all less than 5 and range
from 2.06 to 4.99, and RMSEA values range from .07 to .09.
Comparative fit indices range from .93 to .96 (NNFI) and from .95
to .97 (CFI). Table 7 reports the completely standardized solution
of the factor loadings for the DIPSI facets in the three samples. All
loadings are significant, and only one facet shows a loading below
.50, that is, Withdrawn Traits (loading ⫽ .49), in the sample of
referred children. Reliability analysis of the four underlying factors
demonstrates a high internal consistency with alpha coefficients
from .71 to .86 across the three samples (see Table 7). Factor
correlations are all significant at p ⬍ .01 and show moderate to
substantial correlations in each of the three samples (range from
.26 to .53 for the referred sample, from .36 to .75 for the nonre-
ferred child sample, and from .51 to .80 for the nonreferred
adolescent sample), indicating higher correlations for adolescent
self-ratings compared with maternal ratings of referred and non-
referred children.
Factorial Invariance Across Groups
Tables 8 and 9 present the results of the multisample analyses
exploring the factorial invariance across different age groups and
informants and across referred and nonreferred groups. In a first
analysis (see Table 8), we compared nonreferred with referred
children (mother ratings) and tested a baseline model in which
only the form of the model—the pattern of fixed and nonfixed
parameters—was invariant across groups (Marsh, 1994). In a
second model, the factor loadings were constrained to be equal in
each group. A third model set both factor loadings and unique
variances equal in both groups. The fourth model constrained the
factor loadings, unique variances, and factor correlations to be
equal in each group. The same sequence of tests was repeated for
the multisample analysis of nonreferred children (mother ratings)
versus nonreferred adolescents (self-ratings; see Table 9).
The results reject the hypotheses that the factor loadings, unique
variances, and factor correlations were invariant for both multi-
sample analyses. However, in both analyses,
2
/df ratios were
Table 6
Absolute and Comparative Fit Indices for Maternal Ratings of Nonreferred (N ⫽ 242) and
Referred (N ⫽ 205) Children and Self-Ratings of Nonreferred (N ⫽ 453) Adolescents
Goodness-of-fit indices
Children
(R)
Children
(NR)
Adolescents
(NR)
2
98.64 130.44 239.61
df 48 48 48
2
/df
2.06 2.72 4.99
Root-mean-square error of approximation (Steiger, 1990) .07 .08 .09
Nonnormed fit index (Bentler & Bonnett, 1980) .95 .96 .93
Comparative fit index (Bentler, 1990) .97 .97 .95
Note. The fit index
2
refers to the Satorra–Bentler scaled chi-square. R ⫽ referred; NR ⫽ nonreferred.
650
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
below 5 (2.41 in the first analysis and 3.70 in the second analysis);
RMSEA values were .08 and .09, respectively; CFI values were
.97 and .96, respectively; and NNFI values were .96 and .95,
respectively. These results indicate a basic form of factorial in-
variance, in which the same pattern of zero and nonzero loadings
holds across groups. Hence, the same configuration of loadings on
factors is observed in referred versus nonreferred children (using
the same informant) and in nonreferred children versus nonre-
ferred adolescents (using different informants). This configural
invariance is consistent with the assumption that similar, but not
identical, latent variables can be identified in the different samples
(Widaman & Reise, 1997).
General Discussion
The primary objectives of the present work were the compila-
tion of a taxonomy for the description of personality symptoms
observable in childhood and the examination of the underlying
structure and replicability across different samples. A comprehen-
sive pool of maladaptive trait symptoms was constructed in Study
1 using two different strategies. Descriptors were subgrouped in
item sets that were refined through a series of empirical procedures
in 27 unidimensional and homogenous maladaptive personality
facets. The covariance across facets is hierarchically represented in
a model with at the highest level an internalizing and an
externalizing dimension that further breaks down in three-,
four-, and five-factor solutions. Because the two-factor solution
appeared to be empirically related to the internalizing–
externalizing structure of child psychopathology models (Achen-
bach, 1991), and because the four-dimensional model of personality
pathology in childhood showed a structure similar to that of adults
(Livesley, 1990), the fit of this two–four higher order structure was
further examined in Study 2.
Four-Dimensional Level
The results demonstrate that four higher order factors provide an
acceptable fit to the data and account for a comprehensive amount
Table 7
Completely Standardized Solution of the Four-Factor Model of Personality Symptoms in Maternal Ratings of Nonreferred (N ⫽ 242)
and Referred (N ⫽ 205) Children and in Self-Ratings of Nonreferred (N ⫽ 453) Adolescents
DIPSI marker
Child
(R)
Child
(NR)
Adolescent
(NR)
Child
(R)
Child
(NR)
Adolescent
(NR)
Child
(R)
Child
(NR)
Adolescent
(NR)
Child
(R)
Child
(NR)
Adolescent
(NR)
Disagreeableness
Irritable–Aggressive Traits .82 .81 .81
Dominance–Egocentrism .73 .75 .61
Hyperactive Traits .71 .73 .58
Alpha coefficient .80 .80 .71
Emotional Instability
Dependency .86 .89 .70
Insecure Attachment .88 .86 .57
Anxious Traits .66 .73 .79
Alpha coefficient .84 .86 .74
Compulsivity
Perfectionism .99 1.00 .91
Extreme Achievement Striving .72 .76 .62
Extreme Order .60 .54 .68
Alpha coefficient .78 .79 .76
Introversion
Shyness .79 .90 .80
Paranoid Traits .86 .80 .77
Withdrawn Traits .49 .68 .61
Alpha coefficient .75 .81 .75
Note. All ts for the factor loadings are significant (ts ⬎ 2.58). DIPSI ⫽ Dimensional Personality Symptom Item Pool; R ⫽ referred; NR ⫽ nonreferred.
Table 8
Absolute and Comparative Fit Indices for the Multisample Analysis: Nonreferred (N ⫽ 242)
Versus Referred (N ⫽ 205) Children
Model description
2
df RMSEA CFI NFI NNFI
diff
2
df
diff
p
Model 1: Unrestricted 231.06 96 .08 .97 .95 .96
Model 1: ⌳ invariant 379.71 108 .11 .93 .91 .92 146.29 12 ⬍ .001
Model 2: ⌳⌰invariant 712.45 120 .15 .86 .83 .84 135.13 12 ⬍ .001
Model 3: ⌳⌰⌽invariant 781.47 126 .15 .84 .82 .83 50.74 6 ⬍ .001
Note. The fit index
2
refers to the Satorra–Bentler scaled chi-square. The fit index
diff
2
refers to the
Satorra–Bentler scaled difference chi-square test statistic (Satorra & Bentler, 2001). RMSEA ⫽ root-mean-
square error of approximation; CFI ⫽ comparative fit index; NFI ⫽ normed fit index; NNFI ⫽ nonnormed fit
index; diff ⫽ difference.
651
PERSONALITY PATHOLOGY IN CHILDHOOD
of the covariance of the DIPSI facets, with percentages of ex-
plained variance for the four-factor solution comparable with those
found for the Big Five. The resulting four-factor structure is
conceptually similar to the structures of the DAPP–BQ (Livesley,
1990) and the SNAP (Clark, 1993) taxonomies representing adult
personality pathology, suggesting common dimensions of trait
pathology from childhood and adolescence to adulthood.
The higher order factors were labeled as Disagreeableness,
Emotional Instability, Introversion, and Compulsivity, pointing
toward equivalent factors in the DAPP–BQ and the SNAP (Clark
& Livesley, 2002). The correspondence with the DAPP–BQ is
evident given the similar labels, whereas Emotional Instability,
Introversion, and Compulsivity are conceptually close to Negative
and Positive Affect and Disinhibition in Clark’s (1993) system.
Disagreeableness includes extreme low-end variants of Benevo-
lence (such as Dominance–Egocentrism and Irritable–Aggressive
Traits) but also high-end variants of Extraversion (such as Hyper-
expressive Traits and Hyperactive Traits) and low-end variants of
Conscientiousness (such as Distraction and Disorderliness), and
may hence be considered as somewhat broader than adult concep-
tualizations of this higher order trait. Emotional Instability refers to
both Anxious and Depressive Traits and also includes a Depen-
dency component represented by the facets Insecure Attachment,
Dependency, and Submissiveness. Introversion describes extreme
low-end variants of Extraversion, whereas Compulsivity includes
the high extremes of Conscientiousness Traits.
The nature of these higher order maladaptive trait factors is
reflected in their correlations with adaptive personality traits. Dis-
agreeableness correlates negatively with the HiPIC Benevolence
and Conscientiousness domains and positively with the HiPIC
Extraversion domain. Emotional Instability correlates negatively
with the HiPIC Emotional Stability domain, and the DIPSI Intro-
version and Compulsivity factors relate negatively to HiPIC Ex-
traversion and positively to the HiPIC Conscientiousness dimen-
sions, respectively. The DIPSI Introversion and Compulsivity
dimensions are also related to the Neuroticism component of the
HiPIC. The DIPSI four-factor higher order traits can further be
conceptually related to developmental research on specific mal-
adaptive traits in childhood such as psychopathy (including
callous– unemotional traits, egocentricity, and impulsivity; Forth,
Hare, & Hart, 1990; Frick, O’Brien, Wootton, & McBurnett, 1994;
Lynam, 1997, 1998) or perfectionism (Flett, Hewitt, Boucher,
Davidson, & Munro, 2001). In contrast with psychopathic or
perfectionism measures that represent top-down translations of
adult scales (Lynam et al., 2005), the DIPSI includes maladaptive
traits resulting from a bottom-up perspective and creates the op-
portunity to cover a broader range of maladaptive traits.
The evidence presented at the four-dimensional level leads to
two major conclusions. First, the basic structure of maladaptive
traits and personality pathology is very similar in adulthood and
childhood, underscoring the fact that personality pathology applies
to children and adolescents, hence supporting a developmental
perspective on personality disorders. Second, the DIPSI–HiPIC
associations extend previous results on similar covariation patterns
between categorical personality disorder measures and adaptive
FFM facets in childhood and adulthood (De Clercq & De Fruyt,
2003; De Clercq et al., 2004) to dimensional measures of adaptive
and maladaptive personality. A theoretical comparison with di-
mensional adaptive (NEO Personality Inventory—Revised)–
maladaptive (SNAP) associations in adulthood (Clark, Vorhies, &
McEwen, 2002) further suggests similar DIPSI–HiPIC relation-
ships in childhood.
Two-Dimensional Level
The four dimensions can be integrated in a two-dimensional
structure, with a clear externalizing and an internalizing dimension
at the top of the hierarchy. This two-dimensional structure shows
close correspondence to the two broadband factors distinguished in
Achenbach’s (1991) CBCL system describing childhood problem
behavior but is also very similar to Krueger and Piasecki’s (2002)
hierarchical-spectrum model structuring both Axis I and Axis II
Diagnostic and Statistical Manual of Mental Disorders (American
Psychiatric Association, 1980, 1987, 1994) psychopathologies in
adulthood. Krueger and Piasecki hypothesized that “specific clus-
ters of symptoms cohere to form syndromes, which in turn cohere
in even broader families of disorders or spectra” (p. 487). The
findings from our study contribute to this evidence, with Emo-
tional Instability, Introversion, and Compulsivity merging into the
Internalizing trait factor and Disagreeableness representing the
Externalizing trait factor, suggesting that two broad superfactors of
trait pathology are relevant from childhood to adulthood. More-
over, the conceptual similarities between this superorder maladap-
tive trait structure and the empirical relationships between these
dimensions and the CBCL problem scales suggest a common
framework integrating Axis I-related and Axis II disorders.
Table 9
Absolute and Comparative Fit Indices for the Multisample Analysis: Nonreferred Children (N ⫽
242) Versus Nonreferred Adolescents (N ⫽ 453)
Model description
2
df RMSEA CFI NFI NNFI
diff
2
df
diff
p
Model 0: Unrestricted 355.35 96 .09 .96 .95 .95
Model 1: ⌳ invariant 701.35 108 .13 .91 .90 .89 215.21 12 ⬍ .001
Model 2: ⌳⌰invariant 894.59 120 .14 .88 .87 .87 97.37 12 ⬍ .001
Model 3: ⌳⌰invariant 1,020.34 126 .14 .86 .85 .86 64.77 6 ⬍ .001
Note. The fit index
2
refers to the Satorra–Bentler scaled chi-square. The fit index
diff
2
refers to the
Satorra–Bentler scaled difference chi-square test statistic (Satorra & Bentler, 2001). RMSEA ⫽ root-mean-
square error of approximation; CFI ⫽ comparative fit index; NFI ⫽ normed fit index; NNFI ⫽ nonnormed fit
index; diff ⫽ difference.
652
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
The Two–Four Hierarchy
Although this two-factor perspective has its own merits, the
present study demonstrated that a four-dimensional model, distin-
guishing Disagreeableness, Emotional Instability, Introversion,
and Compulsivity, is replicable in different samples and explains
substantially more variance in trait pathology. The two–four higher
order structure showed an acceptable fit in the sample of nonre-
ferred children and clearly demonstrated how the internalizing
component breaks down into a Negative Emotionality and a Com-
pulsivity component, and Negative Emotionality subsequently un-
folds in a Neuroticism and an Introversion component. The exter-
nalizing dimension is very robust across the two-, the three-, and
the four-dimensional hierarchy and only bifurcates in a Disagree-
able Disinhibition and an Unconscientious Disinhibition compo-
nent at the five-dimensional level. These patterns suggest that
internalizing traits have a more heterogeneous nature relative to
externalizing traits that seem to be more unidimensional.
The multilevel perspectives on the higher order structure of
maladaptive trait differences are not incompatible but rather reflect
differences in the required level of psychopathology integration
and differentiation. A higher level of differentiation has higher
potential for providing additional insight in the nature of a problem
or for providing cues for specific assessment or treatment, whereas
the internalizing– externalizing level enhances researchers’ under-
standing of the comorbidity of mental disorders, integrating both
Axis I-related and Axis II pathology. A multidimensional model
further enables better differentiation and description of pathology
profiles and syndromes, given the fact that the present findings
reveal four DIPSI higher order dimensions that are differentially
related to the CBCL syndrome scales. The present data show, for
example, that the CBCL scale Anxious–Depressed is strongly
related to the Internalizing trait dimension from the two-factor
solution, but the associations with the four-factor level dimensions
show that this syndrome scale is significantly associated with both
Emotional Instability and Introversion and moderately associated
with Compulsivity, providing additional insight into the major
components of this problem scale.
Lower Order Level
Besides similarity at the higher order levels of the trait hierar-
chy, a strong conceptual correspondence between the lower level
structures in childhood and adulthood was observed, although
some differences were apparent. Not all subscales included in the
adult lower order structure are (fully) applicable to denote differ-
ences in children, and the bottom-up approach produced two facets
(Hyperexpressive Traits and Inflexibility) that are not fully repre-
sented as separate entities in the adult maladaptive trait taxono-
mies. Moreover, the DIPSI facets appear to reflect a broader
spectrum of traits (especially the Disagreeableness Traits) and also
facets referring to traits that may blend with Axis I pathology (such
as Separation Anxiety and Insecure Attachment). Potential overlap
with more Axis I-related pathology is inherent in the study of
(maladaptive) traits, especially because individuals’ trait positions
have to be inferred from feelings, thoughts, and behaviors (in
different contexts and across the life course) rather than being
directly observable.
Although our findings underscore the potential of the applied
bottom-up and top-down approaches to provide a comprehensive
account of observable trait differences at young ages, the useful-
ness of these lower level facets remains to be demonstrated in
future research. Some DIPSI facets may be dropped or may require
further revision, relying on empirical evidence about the discrimi-
nant validity for well-defined clinical syndromes or diagnoses. It
may be necessary, for example, to include facets indicating mal-
adaptive sexual behaviors for adolescents (Westen et al., 2003) to
make the DIPSI more comprehensive for describing maladaptive
traits in adolescents.
DIPSI Markers Versus Extended DIPSI
Given the difficulties involved in confirmatory analysis of com-
plex hierarchical trait models (McCrae, Zonderman, Bond, Costa,
& Paunonen, 1996), a basic structure of maladaptive personality in
childhood and adolescence was confirmed using a well-defined but
limited set of DIPSI markers. Selecting the three highest loading
facets from an exploratory analysis, we proposed a hierarchical
model with three markers for each of the four dimensions. Each of
the lower order facets is a valid indicator for one of the four
broadband dimensions across ages (children vs. adolescents), us-
ing different informants (maternal vs. self-reports), and across
groups (clinic referred vs. nonreferred) using the same informant.
However, the results of a cross-group comparison demonstrate
only a basic form of factorial invariance, indicating that configural
invariance across groups with different age ranges, clinical status,
or informant does not imply that the symptom features are iden-
tical for each of these groups. Possible symptom variations may
exist because of age or the presence of psychopathology, and other
markers may emerge from adolescent data. The 12 marker cate-
gories represent the four latent factors of maladaptive traits in
childhood, but the entire maladaptive trait domain—in terms of the
nature and number of facets—may be amended or changed ac-
cording to the available evidence. The basic structure, however, is
likely to be more robust, given its similarity to the structure found
in adulthood.
An Integrative Perspective for DSM–V
Widiger and Clark (2000) and Trull and Durrett (2005) dis-
cussed a number of issues that should be taken into account when
revising the DSM–IV classification system. One alternative route
was focusing on the dimensions underlying comprehensive models
of psychopathology, including personality pathology. Mervielde,
De Clercq, and colleagues (2005) recently argued to also incorpo-
rate a developmental perspective and to broaden the dimensional
view to children and adolescents. The present work is a first step
toward a life span perspective on trait and (Axis I-related) psy-
chopathology. Three central integrative perspectives are outlined
in this work by advocating the two–four hierarchical model. The
first perspective focuses on the issue of multiple psychopatholo-
gies within individuals, with the internalizing–externalizing level
providing a framework to study co-occurrence of common psy-
chopathology and Axis II-related pathology. The second viewpoint
is related to the spectrum idea of normal and abnormal trait
variation. Adaptive and maladaptive trait models show strong
empirical and conceptual relationships, suggesting that they are
quantitatively rather than qualitatively different trait domains. The
final perspective is a developmental one, demonstrating that per-
653
PERSONALITY PATHOLOGY IN CHILDHOOD
sonality pathology can be observed in children and adolescents,
and showing a structure very similar to that of personality pathol-
ogy in adulthood. These joint perspectives show new avenues for
bringing together various dimensional models and literatures under
a unified framework for DSM–V.
Besides conceptual and theoretical interests, the present work is
also interesting from an assessment point of view. Whether the
DIPSI—as an assessment instrument—will capture psychopathol-
ogy in children above instruments focusing on common psycho-
pathology is ultimately an empirical question that goes beyond the
scope of the present work. Similarly, the value of the DIPSI
beyond adaptive trait or temperament models is to be clarified in
future studies. However, considering the way the DIPSI was
constructed, we believe that it provides a unique and comprehen-
sive set of 27 maladaptive trait facets to describe precursors of
adult trait pathology in a more differentiated way than adaptive
models do. Eventually, as suggested by Widiger and Simonsen
(2005), a comprehensive understanding of personality pathology
requires the inclusion of both adaptive and maladaptive trait
measures.
Limitations
The present work has a number of strengths, including the
adoption of multiple perspectives to compile maladaptive traits
and personality symptoms, using different samples of children and
adolescents, and relying on maternal and self-reports. However,
there are also a number of limitations that require attention. First,
although data were collected in four independent samples of vary-
ing ages (children vs. adolescents) and with a different psycholog-
ical status (referred vs. nonreferred), the present findings are
cross-sectional and do not give information on the longitudinal
course and association with dimensions considered relevant to
describe personality symptoms in adulthood. The comparison with
adult taxonomies requires longitudinal research, starting with an
initial DIPSI assessment in childhood, a second one during ado-
lescence, and a follow-up on the DAPP–BQ and/or the SNAP in
adulthood. A second limitation concerns the restricted age range of
the sample of children from the general population (N ⫽ 749)
because of a disproportionately large number of children of 6
years. Nonetheless, the observed relationships are very similar
across the life course. Third, although we aimed to cover the full
range of trait pathology in children, it cannot be guaranteed that all
maladaptive traits and personality symptoms that are specific for
childhood are included in the item pool. In this respect, it is
noteworthy to recall the absence of extreme indicators of the
Imagination factor of the HiPIC in the final item pool. No extreme
low- or high-end descriptors for Imagination were written, because
we did not consider these as indicative of dysfunctioning, for
example, children being very low or very high on fantasy or
creativity. “Having too much fantasy” or “using tools for unin-
tended purposes” could be considered as dysfunctional or mal-
adaptive in adulthood, but this is not necessarily the case in
childhood, given that an active fantasy life is often considered as
a necessary developmental phase. However, we cannot exclude
that addition of such items (using other approaches to define the
maladaptive trait field to be covered) affect the trait hierarchy and
the decomposition in terms of dimensions.
Applied and Clinical Value
Children with a maladaptive personality profile may be partic-
ularly at risk for developing extreme manifestations of personality
traits that characterize personality disorders. Therefore, Caspi et al.
(2003) and Mason et al. (2004) underscored the importance of
early interventions for children with maladaptive patterns of be-
havior. The present study offers a framework for the description of
these early maladaptive trait patterns, but at present we can only
hypothesize about potential relationships with personality disor-
ders developing in late adolescence and adulthood.
In sum, the present framework provides a unique opportunity to
link the structure of childhood personality pathology with litera-
ture on the structure of common mental disorders in adults and
children. Evidence is presented for both the higher order structure
of childhood psychopathology (Achenbach, 1991) and common
psychopathology in adults (Krueger & Piasecki, 2002), that is,
internalizing and externalizing, as well as for the four-factor struc-
ture that characterizes abnormal personality in adults (Livesley,
1990), extending this structure to children. The evidence provided
in the present work suggests that the proposed two–four dimen-
sional hierarchy can bring various dimensional models and liter-
atures together, introducing a life span perspective on Axis
I-related and Axis II pathology under a unified framework for
DSM–V (Widiger et al., 2005).
References
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4 –18
and 1991 profiles. Burlington: University of Vermont, Department of
Psychiatry.
American Psychiatric Association. (1980). Diagnostic and statistical man-
ual of mental disorders (3rd ed.). Washington, DC: Author.
American Psychiatric Association. (1987). Diagnostic and statistical man-
ual of mental disorders (3rd ed., rev.). Washington, DC: Author.
American Psychiatric Association. (1994). Diagnostic and statistical man-
ual of mental disorders (4th ed.). Washington, DC: Author.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psy-
chological Bulletin, 107, 238 –246.
Bentler, P. M., & Bonnett, D. G. (1980). Significance tests and goodness
of fit in the analysis of covariance structures. Psychological Bulletin, 88,
588 – 606.
Bernstein, D. P., Cohen, P., Skodol, A., Bezirganian, S., & Brook, J. S.
(1996). Childhood antecedents of adolescent personality disorders.
American Journal of Psychiatry, 153, 907–913.
Bollen, K. A. (1989). Structural equations with latent variables. New
York: Wiley.
Briggs-Gowan, M. J., Owens, P. L., Schwab-Stone, M. E., Leventhal,
J. M., Leaf, P. J., & Horwitz, S. M. (2003). Persistence of psychiatric
disorders in pediatric settings. Journal of the American Academy of
Child and Adolescent Psychiatry, 42, 1360 –1369.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model
fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation
models (pp. 136–162). Newbury Park, CA: Sage.
Buyst, V., De Fruyt, F., & Mervielde, I. (1994). Parental descriptions of
children’s personality: A five-factor model classification. Psychologica
Belgica, 34, 231–255.
Caspi, A. (2000). The child is father of the man: Personality continuities
from childhood to adulthood. Journal of Personality and Social Psy-
chology, 78, 158–172.
Caspi, A., Harrington, H., Milne, B., Arnell, J. W., Theodore, R. F., &
Moffitt, T. E. (2003). Children’s behavioral styles at age 3 are linked to
654
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
their adult personality traits at age 26. Journal of Personality, 71,
475–513.
Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality develop-
ment: Stability and change. Annual Review of Psychology, 56, 453– 484.
Clark, L. A. (1990). Toward a consensual set of symptom clusters for
assessment of personality disorder. In J. N. Butcher & C. D. Spielberger
(Eds.), Advances in personality assessment (Vol. 8, pp. 243–266). Hills-
dale, NJ: Erlbaum.
Clark, L. A. (1993). Manual for the Schedule for Nonadaptive and Adap-
tive Personality (SNAP). Minneapolis: University of Minnesota Press.
Clark, L. A., & Livesley, J. W. (2002). Two approaches to identifying the
dimensions of personality disorder: Convergence on the five-factor
model. In P. T. Costa Jr. & T. A. Widiger (Eds.), Personality disorders
and the five-factor model of personality (2nd ed., pp. 161–176). Wash-
ington, DC: American Psychological Association.
Clark, L. A., Livesley, W. J., Schroeder, M. L., & Irish, S. L. (1996).
Convergence of two systems for assessing personality disorder. Psycho-
logical Assessment, 8, 294 –303.
Clark, L. A., Vorhies, L., & McEwen, J. L. (1994). Personality disorder
symptomatology from the five-factor model perspective. In P. T. Costa
Jr. & T. A. Widiger (Eds.), Personality disorders and the five-factor
model of personality (pp. 95–116). Washington, DC: American Psycho-
logical Association.
Clark, L. A., Vorhies, L., & McEwen, J. L. (2002). Personality disorder
symptomatology from the five-factor model perspective. In P. T. Costa
& T. A. Widiger (Eds.), Personality disorders and the five-factor model
of personality (2nd ed., pp. 125–147). Washington, DC: American
Psychological Association.
Cohen, P., Cohen, J., Kasen, S., Velez, C. N., Hartmark, C., Johnson, J., et
al. (1993). An epidemiological study of disorders in late childhood and
adolescence: Age- and gender-specific prevalence. Journal of Child
Psychology and Psychiatry, 34, 851–867.
Costa, P. T., & McCrae, R. R. (1992). Professional Manual: Revised NEO
Personality Inventory (NEO-PI-R) and NEO Five-Factor-Inventory
(NEO-FFI). Odessa, FL: Psychological Assessment Resources.
Costa, P. T., & Widiger, T. A. (1994). Personality disorders and the
five-factor model of personality. Washington, DC: American Psycho-
logical Association.
Costello, E. J., Egger, H., & Angold, A. (2005). 10-year research update
review: The epidemiology of child and adolescent psychiatric disorders:
I. Methods and public health burden. Journal of the American Academy
of Child and Adolescent Psychiatry, 44, 972–986.
De Clercq, B., & De Fruyt, F. (2003). Personality disorder symptoms in
adolescence: A five-factor model perspective. Journal of Personality
Disorders, 17, 269–292.
De Clercq, B., De Fruyt, F., & Mervielde, I. (2003). Construction of the
Dimensional Personality Symptom Item Pool. Unpublished manuscript,
Ghent University.
De Clercq, B., De Fruyt, F., & Van Leeuwen, K. (2004). A “little five”
lexically based perspective on personality disorder symptoms in adoles-
cence. Journal of Personality Disorders, 18, 479 – 499.
De Fruyt, F., De Clercq, B., Van de Wiele, L., & Van Heeringen, C.
(2006). The validity of Cloninger’s psychobiological model versus the
five-factor model to predict DSM–IV personality disorders in a hetero-
geneous psychiatric sample: Domain facet and residualized facet de-
scriptions. Journal of Personality, 74, 479 –510.
De Fruyt, F., Mervielde, I., Hoekstra, H. A., & Rolland, J.-P. (2000).
Assessing adolescents’ personality with the NEO PI-R. Assessment, 7,
329 –345.
Digman, J. M. (1989). Five robust trait dimensions: Development, stability,
and utility. Journal of Personality, 57, 195–214.
Digman, J. M., & Inouye, J. (1986). Further specification of the five robust
factors of personality. Journal of Personality and Social Psychology, 50,
116 –123.
Dyce, J. A., & O’Connor, B. P. (1998). Personality disorders and the
five-factor model: A test of facet-level predictions. Journal of Person-
ality Disorders, 12, 31– 45.
Finch, J. F., & West, S. G. (1997). The investigation of personality
structure: Statistical models. Journal of Research in Personality, 31,
439 – 485.
First, M. B., Gibbon, M., Spitzer, R. L., & Williams, J. B. W., & Benjamin,
L. S. (1997). Structured Clinical Interview for DSM–IV Axis II Person-
ality Disorders (SCID-II). Washington, DC: American Psychiatric Press.
Flett, G. L., Hewitt, P. L., Boucher, D., Davidson, L., & Munro, Y. (2001).
The Child–Adolescent Perfectionism Scale: Development, validation,
and association with adjustment. Manuscript submitted for publication.
Forth, A. E., Hare, R. D., & Hart, S. D. (1990). Assessment of psychopathy
in male young offenders. Psychological Assessment, 2, 342–344.
Frick, P. J., O’Brien, B. S., Wootton, J. M., & McBurnett, K. (1994).
Psychopathy and conduct problems in children. Journal of Abnormal
Psychology, 103, 700–707.
Gerbing, D. W., & Hamilton, J. G. (1996). Viability of exploratory factor
analysis as a precursor to confirmatory factor analysis. Structural Equa-
tion Modeling, 3, 62–72.
Goodwin, R. D., Fergusson, D. M., & Horwood, L. J. (2004). Early
anxious/withdrawn behaviours predict later internalising disorders.
Journal of Child Psychology and Psychiatry, 45, 874– 883.
Haigler, E. D., & Widiger, T. A. (2001). Experimental manipulation of
NEO-PI-R items. Journal of Personality Assessment, 77, 339–358.
John, O. P., Caspi, A., Robins, R. W., Moffit, T. E., & Stouthamer-Loeber,
M. (1994). The “little five”: Exploring the nomological network of the
Five-Factor Model of personality in adolescent boys. Child Develop-
ment, 65, 160–178.
Jo¨reskog, K. G. (1993). Testing structural equation models. In K. A. Bollen
& J. S. Long (Eds.), Testing structural equation models (pp. 294 –316).
Newbury Park, CA: Sage.
Jo¨reskog, K. G., & So¨rbom, D. (1989). LISREL 7: A guide to the program
and applications (2nd ed.) [Computer software manual]. Chicago: SPSS.
Jo¨reskog, K. G., & So¨rbom, D. (2005). LISREL 8.72 [Computer software].
Lincolnwood, IL: Scientific Software International.
Kasen, S., Cohen, P., Skodol, A. E., Johnson, J. G., & Brook. J. S. (1999).
Influence of child and adolescent psychiatric disorders on young adult
personality disorder. American Journal of Psychiatry, 156, 1529 –1535.
Kelloway, E. K. (1996). Common practices in structural equation model-
ing. In C. L. Cooper & I. Robertson (Eds.), International review of
industrial and organizational psychology (pp. 141–180). Chichester,
United Kingdom: Wiley.
Klein, D. N., & Schwartz, J. E. (2002). The relation between depressive
symptoms and borderline personality disorder features over time in
dysthymic disorder. Journal of Personality Disorders, 16, 523–535.
Kohnstamm, G. A., Halverson, C. F., Jr., Mervielde, I., & Havill, V. L.
(Eds.). (1998). Parental descriptions of child personality: Developmen-
tal antecedents of the big five? Hillsdale, NJ: Erlbaum.
Krueger, R. F., & Piasecki, T. M. (2002). Toward a dimensional and
psychometrically informed approach to conceptualizing psychopathol-
ogy. Behaviour Research and Therapy, 40, 485– 499.
Krueger, R. F., & Tackett, J. L. (2003). Personality and psychopathology:
Working toward the bigger picture. Journal of Personality Disorders,
17, 109 –128.
Lamb, M. E., Chuang, S. S., Wessels, H., Broberg, A. G., & Hwang, C. P.
(2002). Emergence and construct validation of the big five factors in
early childhood: A longitudinal analysis of their ontogeny in Sweden.
Child Development, 73, 1517–1524.
Lemery, K. S., Essex, M. J., & Smider, N. A. (2002). Revealing the relation
between temperament and behavior problem symptoms by eliminating
measurement confounding: Expert ratings and factor analyses. Child
Development, 73, 867–882.
Lewinsohn, P. M., Rohde, P., Seeley, J. R., & Klein, D. N. (1997). Axis II
655
PERSONALITY PATHOLOGY IN CHILDHOOD
psychopathology as a function of Axis I disorders in childhood and
adolescence. Journal of the American Academy of Child and Adolescent
Psychiatry, 36, 1752–1759.
Livesley, W. J. (1986). Traits and behavioral prototypes of personality
disorder. American Journal of Psychiatry, 143, 728 –732.
Livesley, W. J. (1990). Dimensional Assessment of Personality Path-
ology—Basic Questionnaire. Unpublished manuscript, University of
British Columbia, Vancouver, British Columbia, Canada.
Livesley, W. J. (2003). Introduction to the special feature on personality,
personality disorder, and psychopathology. Journal of Personality Dis-
orders, 17, 87–89.
Livesley, W. J., Jang, K. L., & Vernon, P. A. (1998). Phenotypic and
genetic structure of traits delineating personality disorder. Archives of
General Psychiatry, 55, 941–948.
Livesley, W. J., Schroeder, M. L., Jackson, D. N., & Jang, K. L. (1994).
Categorical distinctions in the study of personality disorder: Implications
for classification. Journal of Abnormal Psychology, 103, 6–17.
Lynam, D. R. (1997). Childhood psychopathy: Capturing the fledgling
psychopath in a nomological net. Journal of Abnormal Psychology, 106,
209 –234.
Lynam, D. R. (1998). Early identification of the fledgling psychopath:
Locating the psychopathic child in the current nomenclature. Journal of
Abnormal Psychology, 107, 425– 438.
Lynam, D. R., Caspi, A., Moffitt, T. E., Raine, A., Loeber, R., &
Stouthamer-Loeber, M. (2005). Adolescent psychopathy and the Big
Five: Results from two samples. Journal of Abnormal Child Psychology,
33, 431– 443.
Markon, K. E., Krueger, R. F., & Watson, D. (2005). Delineating the
structure of normal and abnormal personality: An integrative hierarchi-
cal approach. Journal of Personality and Social Psychology, 88, 139 –
157.
Marsh, H. W. (1994). Confirmatory factor analysis models of factorial
invariance: A multifaceted approach. Structural Equation Modeling, 1,
5–34.
Mason, W. A., Kosterman, R., Hawkins, J. D., Herrenkohl, T. I., Lengua,
L. J., & McCauley, E. (2004). Predicting depression, social phobia, and
violence in early adulthood from childhood behavior problems. Journal
of the American Academy of Child and Adolescent Psychiatry, 43,
307–315.
McCrae, R. R., Zonderman, A. B., Bond, M. H., Costa, P. T., & Paunonen,
S. V. (1996). Evaluating replicability of factors in the revised NEO
Personality Inventory: Confirmatory factor analysis versus procrustes
rotation. Journal of Personality and Social Psychology, 70, 552–566.
McDonald, R. P., & Ringo Ho, M.-H. (2002). Principles and practice in
reporting structural equation analyses. Psychological Methods, 1, 64 –
82.
Mervielde, I., De Clercq, B., De Fruyt, F., & Van Leeuwen, K. (2005).
Temperament, personality, and developmental psychopathology as
childhood antecedents of personality disorders. Journal of Personality
Disorders, 19, 171–201.
Mervielde, I., & De Fruyt, F. (1999). Construction of the Hierarchical
Personality Inventory for Children (HiPIC). In I. Mervielde, I. Deary, F.
De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe (pp.
107–127). Tilburg, the Netherlands: Tilburg University Press.
Mervielde, I., & De Fruyt, F. (2002). Assessing children’s traits with the
Hierarchical Personality Inventory for Children. In B. De Raad & M.
Perugini (Eds.), Big Five assessment (pp. 129 –146). Seattle, WA:
Hogrefe & Huber.
Mervielde, I., De Fruyt, F., & De Clercq, B. (2005). Manual of the
Hierarchical Personality Inventory for Children (HiPIC). Unpublished
manuscript, Ghent University, Ghent, Belgium.
Morey, L. C., Gunderson, M. D., Quigley, B. D., Shea, M. T., Skodol,
A. E., McGlashan, T. H., et al. (2002). The representation of borderline,
avoidant, obsessive– compulsive, and schizotypal personality disorders
by the five-factor model. Journal of Personality Disorders, 16, 215–234.
O’Connor, B. P. (2005). A search for consensus on the dimensional
structure of personality disorders. Journal of Clinical Psychology, 61,
323–345.
Pine, D. S., Cohen, E., Cohen, P., & Brook, J. (1999). Adolescent depres-
sive symptoms as predictors of adult depression: Moodiness or mood
disorder? American Journal of Psychiatry, 156, 133–135.
Poulton, R., Caspi, A., Moffitt, T. E., Cannon, M., Murray, R., & Har-
rington, H. (2000). Children’s self-reported psychotic symptoms and
adult schizophreniform disorder—A 15-year longitudinal study. Ar-
chives of General Psychiatry, 57, 1053–1058.
Ramklint, M., von Knorring, A., von Knorring, L., & Ekselius, L. (2003).
Child and adolescent psychiatric disorders predicting adult personality
disorder: A follow-up study. Nordic Journal of Psychiatry, 57, 23–28.
Roza, S. J., Hofstra, M. B., van der Ende, J., & Verhulst, F. C. (2003).
Stable prediction of mood and anxiety disorders based on behavioral and
emotional problems in childhood: A 14-year follow-up during child-
hood, adolescence, and young adulthood. American Journal of Psychi-
atry, 160, 2116–2121.
Samuel, D. B., & Widiger, T. A. (2004). Clinicians’ personality descrip-
tions of prototypic personality disorders. Journal of Personality Disor-
ders, 18, 286–308.
Satorra, A., & Bentler, P. M. (2001). A scaled difference chi-square test
statistic for moment structure analysis. Psychometrika, 66, 507–514.
Saulsman, L. M., & Page, A. C. (2004). The five-factor model and
personality disorder empirical literature: A meta-analytic review. Clin-
ical Psychology Review, 23, 1055–1085.
Schermelleh-Engel, K., Mossbrugger, H., & Mu¨ller, H. (2003). Evaluating
the fit of structural equation models: Tests of significance and descrip-
tive goodness-of-fit measures. Methods of Psychological Research On-
line, 2, 23–74.
Schotte, C., & De Doncker, D. (1996). ADP-IV Questionnaire: Manual
and Norms. Antwerp, Belgium: University Hospital Antwerp.
Schotte, C. K., De Doncker, D. A., Dmitruk, D., Van Mulders, I.,
D’Haenen, H., & Cosyns, P. (2004). The ADP-IV questionnaire: Dif-
ferential validity and concordance with the semi-structured interview.
Journal of Personality Disorders, 18, 405– 419.
Schotte, C. K., De Doncker, D., Vankerckhoven, C., Vertommen, H., &
Cosyns, P. (1998). Self-report assessment of the DSM–IV personality
disorders. Measurement of trait and distress characteristics: The ADP-
IV. Psychological Medicine, 28, 1179–1188.
Schroeder, M. L., Wormworth, J. A., & Livesley, W. J. (1992). Dimensions
of personality disorder and their relationship to the Big Five dimensions
of personality. Psychological Assessment, 4, 47–53.
Shiner, R. (2000). Linking childhood personality with adaptation: Evi-
dence for continuity and change across time into late adolescence.
Journal of Personality and Social Psychology, 78, 310–325.
Shiner, R., & Caspi, A. (2003). Personality differences in childhood and
adolescence: Measurement, development, and consequences. Journal of
Child Psychology and Psychiatry, 44, 2–32.
Shiner, R. L., & Masten, A. S. (2002). Transactional links between per-
sonality and adaptation from childhood through adulthood. Journal of
Research in Personality, 36, 580–588.
Shiner, R. L., Masten, A. S., & Roberts, J. M. (2003). Childhood person-
ality foreshadows adult personality and life outcomes two decades later.
Journal of Personality, 71, 1145–1170.
Skodol, A. E., Rosnick, L., Kellman, H. D., Oldham, J., & Hyler, S. E.
(1988). Validating structured DSM–III–R personality disorder assess-
ments with longitudinal data. American Journal of Psychiatry, 145,
1297–1299.
Steiger, J. H. (1990). Structural model evaluation and modification: An
interval estimation approach. Multivariate Behavioral Research, 25,
173–180.
656
DE CLERCQ, DE FRUYT, VAN LEEUWEN, AND MERVIELDE
Stevens, J. (1996). Exploratory and confirmatory factor analysis. In J.
Stevens (Ed.), Applied multivariate statistics for the social sciences (3rd
ed., pp. 362–428). Mahwah, NJ: Erlbaum.
Tillfors, M., Furmark, T., Ekselius, L., & Fredrikson, M. (2004). Social
phobia and avoidant personality disorder: One spectrum disorder? Nor-
dic Journal of Psychiatry, 58, 147–152.
Trull, T. J., & Durrett, C. A. (2005). Categorical and dimensional models
of personality disorder. Annual Review of Clinical Psychology, 1, 1–26.
Trull, T. J., & Widiger, T. A. (1997). Structured Interview for the Five-
Factor Model of Personality (SIFFM): Professional manual. Odessa,
FL: Psychological Assessment Resources.
Trull, T. J., Widiger, T. A., & Burr, R. (2001). A structured interview for
the assessment of the five-factor model of personality: Facet level
relations to the Axis II personality disorders. Journal of Personality, 69,
175–198.
Tyrer, P., Gunderson, J., Lyons, M., & Tohen, M. (1997). Special feature:
Extent of comorbidity between mental state and personality disorders.
Journal of Personality Disorders, 11, 242–259.
Verhulst, F. C., Van der Ende, J., & Koot, H. M. (1996). Handleiding voor
de CBCL/4–18. [Manual of the CBCL/4 –18.]. Rotterdam, the Nether-
lands: Erasmus Universiteit, Afdeling Kinder-en Jeugdpsychiatrie.
Westen, D., & Shedler, J. (1999a). Revising and assessing Axis II: Part I:
Developing a clinically and empirically valid assessment method. Amer-
ican Journal of Psychiatry, 156, 258 –272.
Westen, D., & Shedler, J. (1999b). Revising and assessing Axis II: Part II:
Toward an empirically based and clinically useful classification of
personality disorders. American Journal of Psychiatry, 156, 273–285.
Westen, D., Shedler, J., Durrett, C., Glass, S., & Martens, A. (2003).
Personality diagnoses in adolescence: DSM–IV Axis II diagnoses and an
empirically derived alternative. American Journal of Psychiatry, 160,
952–966.
Widaman, K. (1993). Common factor analysis versus principal component
analysis: Differential bias in representing model parameters? Multivar-
iate Behavioral Research, 28, 263–311.
Widaman, K. F., & Reise, S. P. (1997). Exploring the measurement
invariance of psychological instruments: Applications in the substance
use domain. In K. Bryant, M. Windle, & S. West (Eds.), The science of
prevention: Methodological advances from alcohol and substance abuse
research (pp. 281–324). Washington, DC: American Psychological
Association.
Widiger, T. A. (2003). Personality disorder and Axis I psychopathology:
The problematic boundary of Axis I and II. Journal of Personality
Disorders, 17, 90–108.
Widiger, T. A., & Clark, L. A. (2000). Toward DSM–V and the classifi-
cation of psychopathology. Psychological Bulletin, 126, 946 –963.
Widiger, T. A., & Costa, P. T., Jr. (2002). Personality disorders and the
five-factor model of personality (2nd ed.). Washington, DC: American
Psychological Association.
Widiger, T. A., & Frances, A. J. (2002). Toward a dimensional model of
personality disorders. In P. T. Costa Jr. & T. A. Widiger (Eds.), Per-
sonality disorders and the five-factor model of personality (2nd ed., pp.
23– 44). Washington, DC: American Psychological Association.
Widiger, T. A., & Simonsen, E. (2005). Alternative dimensional models of
personality disorder: Finding a common ground. Journal of Personality
Disorders, 19, 110–130.
Widiger, T. A., Simonsen, E., Krueger, R., Livesley, J., & Verheul, R.
(2005). Personality disorder research agenda for the DSM–V. Journal of
Personality Disorders, 19, 315–338.
Widiger, T. A., Trull, T. J., Clarkin, J. F., Sanderson, C., & Costa, P. T.
(2002). A description of the DSM–IV personality disorders with the
five-factor model of personality. In P. T. Costa Jr. & T. A. Widiger
(Eds.), Personality disorders and the five-factor model of personality
(2nd ed., pp. 89–99). Washington, DC: American Psychological
Association.
Received May 24, 2005
Revision received December 2, 2005
Accepted March 13, 2006 䡲
657
PERSONALITY PATHOLOGY IN CHILDHOOD
