Article

Profile of children investigated for sexual abuse: Association with psychopathology symptoms and services

School of Social Work, University of North Carolina, Chapel Hill, NC, USA.
American Journal of Orthopsychiatry (Impact Factor: 1.36). 11/2006; 76(4):468-81. DOI: 10.1037/0002-9432.76.4.468
Source: PubMed

ABSTRACT

Sexually abused children may have poor mental health because of their victimization as well as preexisting or co-occurring family problems. However, few studies consider psychopathology in relation to both abuse and other family experiences. This study uses data from the National Survey of Child and Adolescent Well-Being (NSCAW) to create latent subgroups of 553 children investigated for sexual abuse. The study investigates children's psychological symptoms and child welfare service (CWS) patterns to understand how children's needs relate to mental health services. Analyses were conducted by child age: 3-7, 8-11, and 12-14. Factor mixture modeling and regression analyses were used. Results show meaningful subgroups of children that relate to different symptom patterns. Among 3- to 7-year-olds, behavioral symptoms are associated with caregiver domestic violence and mental illness. Among 8- to 11-year-olds, depressive symptoms are associated with severe abuse and multiple family problems, whereas posttraumatic stress is associated with chronic, unresolved abuse. Although many children received mental health services, services are not well matched to children's needs--the substantiation status of the abuse explains services. Implications for CWS and mental health services are discussed.

Full-text

Available from: Mimi V. Chapman
Profile of Children Investigated for Sexual Abuse: Association With
Psychopathology Symptoms and Services
Julie S. McCrae, MA, MSW, Mimi V. Chapman, PhD, and Sharon L. Christ, MS
University of North Carolina at Chapel Hill
Sexually abused children may have poor mental health because of their victimization as well as
preexisting or co-occurring family problems. However, few studies consider psychopathology in relation
to both abuse and other family experiences. This study uses data from the National Survey of Child and
Adolescent Well-Being (NSCAW) to create latent subgroups of 553 children investigated for sexual
abuse. The study investigates children’s psychological symptoms and child welfare service (CWS)
patterns to understand how children’s needs relate to mental health services. Analyses were conducted
by child age: 3–7, 8 –11, and 12–14. Factor mixture modeling and regression analyses were used. Results
show meaningful subgroups of children that relate to different symptom patterns. Among 3- to 7-year-
olds, behavioral symptoms are associated with caregiver domestic violence and mental illness. Among
8- to 11-year-olds, depressive symptoms are associated with severe abuse and multiple family problems,
whereas posttraumatic stress is associated with chronic, unresolved abuse. Although many children
received mental health services, services are not well matched to children’s needs—the substantiation
status of the abuse explains services. Implications for CWS and mental health services are discussed.
Keywords: sexual abuse, NSCAW, latent profile analysis, child welfare services
Much of what we know about child sexual abuse (CSA) and
psychopathology has come from studies that focus on the charac-
teristics of the abuse as the primary determinant of subsequent
problems. Increasingly, research shows that maltreatment types
often co-occur or change over time (Jonson-Reid, Drake, Chung,
& Way, 2003; Lau et al., 2005; Manly, Cicchetti, & Barnett, 1994).
Sexually abused children are likely to be exposed to other serious
family problems, such as domestic violence (DV), physical abuse,
and substance abuse (Dong, Anda, Dube, Giles, & Felitti, 2003;
Kellogg & Menard, 2003). The “main effect” model of CSA on
child outcomes has been challenged in favor of research that is
attuned to the reality of multitype maltreatment and cumulative
risk (Turner, Finkelhor, & Ormrod, 2006) and that seeks to better
explain children’s developmental trajectories (e.g., Thornberry,
Ireland, & Smith, 2001). Although research links sexual abuse to
psychopathology, the complexity of children’s lives is such that
using sexual abuse as a sole criterion of service referrals might not
optimally direct clinical and child welfare resources. Evidence that
up to two thirds of sexually abused children are asymptomatic
(Kendall-Tackett, Williams, & Finkelhor, 1993), yet are provided
mental health and other services at rates greater than other mal-
treated children (Garland, Landsverk, Hough, & Ellis-MacLeod,
1996), suggests that children’s psychological needs, service pat-
terns, and needs arising from additional stressors following sexual
abuse need further study.
Efforts to better understand the children’s psychological trajec-
tories in relation to multitype maltreatment and cumulative or
co-occurring family problems coincide with efforts to provide
more tailored and developmentally sensitive child welfare services
(Berrick, Needell, Barth, & Jonson-Reid, 1998; Waldfogel, 1998).
New federal standards hold child welfare agencies accountable to
address child well-being in addition to meeting child safety and
permanency goals. Results of the first Child and Family Service
Reviews in 2001 and 2002 are bleak; 30 of 32 states with com-
pleted reviews were judged in need of improvement with regard to
meeting children’s mental health service needs (USDHHS [U.S.
Department of Health and Human Services], 2002). Child welfare
services (CWS) that have traditionally focused limited resources
on parents (Wolfe, 1999) will be challenged to meet these new
standards. This study endeavors to contribute to the literature about
psychopathology and sexual abuse by using national survey data to
examine the collective influence of abuse and family problems on
children’s psychopathology symptoms and children’s subsequent
referral to, or receipt of, mental health services.
Sexually abused children are reportedly referred for mental
health services and CWS more often than other children, perhaps
because CSA is a frequently assumed precursor to psychopathol-
ogy. Data concerning 662 children in foster care showed that
sexually abused children are referred to mental health services
Julie S. McCrae, MA, MSW, and Mimi V. Chapman, PhD, School of
Social Work, University of North Carolina at Chapel Hill; Sharon L.
Christ, MS, Odum Institute, University of North Carolina at Chapel Hill.
This research was supported by a doctoral dissertation fellowship to
Julie S. McCrae from the Children’s Bureau of the U.S. Department of
Health and Human Services (USDHHS), Administration on Children and
Families (Award 90CA1718/01). The National Survey of Child and Ado-
lescent Well-Being is funded under a contract from the Administration on
Children and Families, USDHHS. We thank Richard P. Barth and Shen-
yang Guo for their substantive and analytic contributions to the project.
Points of view or opinions in this article are those of the authors and do
not necessarily represent the official position or policy of the sponsors.
For reprints and correspondence: Julie S. McCrae, MA, MSW, School
of Social Work, University of North Carolina, Chapel Hill, NC 27599-
3550. E-mail: jsmccrae@email.unc.edu
American Journal of Orthopsychiatry Copyright 2006 by the American Psychological Association
2006, Vol. 76, No. 4, 468 481 0002-9432/06/$12.00 DOI: 10.1037/0002-9432.76.4.468
468
Page 1
more often than other children, regardless of exhibited symptoms
(Garland, Landsverk, Hough, & Ellis-MacLeod, 1996). Greater
rates of child welfare case opening were found among investiga-
tions of sexual abuse, a pattern that was linked to the perception
among child welfare workers that CSA is particularly harmful to
children (McCrae & Barth, 2004). However, sexually abused chil-
dren in this study were not different from other children with
regard to their reported exposure to other family problems such as
substance abuse, prior maltreatment, and domestic violence—
factors that also pose threats to their mental health.
Certainly, the perception that sexual abuse is a precursor to
psychopathology has considerable merit. Over 40 years of research
supports the link between CSA and psychopathology, including
posttraumatic stress disorder (PTSD), depression, and behavior
problems (Avery, Massat, & Lundy, 2000; Beitchman, Zucker,
Hood, DaCosta, & Akman, 1991; Kendall-Tackett et al., 1993;
McLeer et al., 1998; Romano & De Luca, 1999). However, many
children do not show signs of mental distress, and researchers
agree that no specific psychological or behavioral “syndrome”
follows sexual abuse (Beichtman et al., 1991; Kendall-Tackett et
al., 1993). Most children show recovery from symptoms during the
first 12 to 18 months (Finkelhor et al., 1990; Kendall-Tackett et al.,
1993). The phenomenon of “sleeper effects”—in which asymp-
tomatic children later become symptomatic (Fergusson & Mullen,
1999; Finkelhor & Berliner, 1995; Putnam, 1996)—further com-
plicates the design of services, because it is unclear which children
need services and when. A strong emotional or behavioral reaction
after CSA does not always correspond with long-term problems,
and the apparent absence of symptoms does not guarantee a path
free from mental disorders. Data showing that sexual abuse is
predictive of adult mental disorders, regardless of individuals’
exposure to other childhood adversities (Molnar, Buka, & Kessler,
2001), lends support to the idea that all CSA-exposed children
ought to be provided mental health services.
Research also shows, however, that children’s sexual abuse
experiences vary greatly and data from several studies suggest that
many children exposed to sexual abuse may not be at high risk for
developing psychological problems. Characteristics of sexual
abuse that portend poor outcomes include abuse of longer duration,
that involves penetration or force, and a perpetrator who is close to
the child (Kendall-Tackett et al., 1993). Indeed, one study of 80
nonclinically referred children showed that CSA-related and de-
mographic factors accounted for greater than one half of the
variance in children’s global functioning and predicted PTSD in
over 80% of children (Ruggiero et al., 2000), yet over one half of
children (60%) in this study had reportedly experienced CSA on
one occasion. A review of eight studies describing abuse charac-
teristics showed that roughly 10% of CSA incidents involved close
family members, 50% involved acquaintances, and just 5% to 10%
of CSA incidents involved the most severe level of abuse defined
as intercourse (Fergusson & Mullen, 1999). The predominance of
children who have experienced relatively less severe CSA prompts
questions about the utility of universal or preventive mental health
services for CSA-exposed children. If abuse dimensions such as
duration and perpetrator closeness are the primary determinants of
children’s psychological course, many children may not be at high
risk, particularly in situations wherein there are strong emotional
supports available to children following the abuse, such as from a
nonoffending caregiver or other adult (Cohen & Mannarino,
1996).
There may be other factors that influence, or that may even more
fully explain in some situations, children’s psychopathology risk.
Sexual abuse accompanied by neglect is associated with higher
levels of internalizing behavior problems (Bolger & Patterson,
2001). Sexual abuse, along with physical and emotional abuse,
family disruption, and poverty predicted the most clinical depres-
sion and anxiety among 290 female adolescents in Canada (Bagley
& Malick, 2000). Data from the National Survey of Adolescents
show that depressed adolescents who experienced both sexual and
physical abuse exhibit the most symptoms (Danielson, de Arel-
lano, Kilpatrick, Saunders, & Resnick, 2005) and comorbid ado-
lescent psychopathology is associated with several types of inter-
personal violence experiences (Kilpatrick et al., 2003). CWS that
are configured to better meet children’s mental health service
needs will need to be cognizant of multiple factors that put them at
risk for developing problems. In addition, children’s psychological
needs may change over time in relation to maltreatment and related
family problems. Investigators who examined the timing of mal-
treatment, children’s exposure to multiple family problems, and
their psychological symptoms, for example, found that sexually
abused children had worse outcomes that were attributable to their
CSA experiences at the onset, but which were attributable to other
family problems 6 months later (Paradise, Rose, Sleeper, &
Nathanson, 1994). Children’s problems at 6 months in this study
were predicted by factors such as having prior reports, poor family
integration, lower maternal education, and maternal psychiatric
problems. Studies such as these are a first step toward better
understanding children’s psychological course following sexual
abuse that may aid the design of CWS and mental health services
that are tailored to meet children’s individual needs.
Current research is limited not only because many studies have
focused on the characteristics of the CSA that best predict chil-
dren’s problems, but also because much of what we currently
know is based on adult retrospective reports or reports of select
subgroups of children, such as those with substantiated abuse or
clinical-level problems (e.g., Avery, Massat, & Lundy, 2000;
Molnar et al., 2001; Garland et al., 1996). CWS and mental health
service resources that are sufficiently targeted in cases of sexual
abuse are complicated by the paucity of information about the
larger group of children investigated for sexual abuse. For the first
time, national data are available through the National Survey of
Child and Adolescent Well-Being (NSCAW). This study allows
for a comprehensive view of the circumstances of CSA that is
brought to the attention of CWS. This study aims to expand our
knowledge of the needs of sexually abused children by identifying
meaningful subgroups of children who may be exhibiting, or at
risk for, psychopathology because of their total abuse and family
experiences. Children in three age groups are analyzed (ages 3–7
years, 8 –11 years, and 12–14 years), in the interest of conducting
child welfare research that is developmentally sensitive (Berrick,
Needell, Barth, & Jonson-Reid, 1998). Children’s symptoms are
viewed from a cumulative risk perspective of how children de-
velop mental disorders, a strategy that likely reflects the day-to-
day circumstances faced by many clinicians and child welfare
workers, wherein children’s problems are not readily attributable
to a single cause.
469
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 2
Method
The NSCAW study involves a sample of 5,501 children, from birth to 14
years of age, investigated for maltreatment between October 1999 and
December 2000. Of the 5,501 children, 12% (n 597) were investigated
for sexual abuse as the most serious maltreatment type. Current analyses
include 553 of these children, ages 3–14 years. Very young children (from
birth to 2 years of age) were excluded because they are a small group
(unweighted n 44) for which standardized measures of psychopathology
or problem behavior are not available. Data are from baseline interviews
with children, their primary caregiver, and child welfare workers.
Survey Design
Children were selected into the NSCAW study using a two-stage strat-
ified sample design. The first stage involved selecting 92 primary dampling
units (PSUs; predominantly county child welfare agencies) in 36 states.
The second stage involved selecting children, from birth to 14 years of age,
from lists or files of children who were investigated for maltreatment
within the sample PSUs. Eligible children were stratified into eight do-
mains according to their age, service status (e.g., open for CWS, foster
care), and maltreatment type, and a random sample of children was chosen
from each domain. This procedure allowed for oversampling of infants,
children reported for sexual abuse, and children receiving ongoing CWS.
Sampling weights were calculated to adjust for unequal selection of sub-
jects into the study, initial nonresponse, and undercoverage of unsubstan-
tiated cases and also to allow statistical inference to the national level of
children investigated for maltreatment. Additional detail about the
NSCAW design and sampling is available (see NSCAW Research Group,
2002).
Measures
Measures are grouped according to variables used to form the latent
class profiles (abuse and family characteristics), measures to assess chil-
dren’s psychological symptoms, and services. Latent classes were formed
using five dichotomous abuse-related variables (duration, severity, parent
perpetrator, relative perpetrator, and maltreatment co-occurrence) and five
family problem variables (prior reports, low social support, substance
abuse, mental illness [MI], and DV).
Abuse characteristics. A modified version of the Maltreatment Clas-
sification System (MCS; Barnett, Manly, & Cicchetti, 1993) was used to
describe the investigated abuse. Child welfare workers reported the dura-
tion and frequency of the CSA, the alleged perpetrator, abuse severity, and
other reported maltreatment. Two variables were created to describe the
perpetrator: parent or stepparent and other relative (both yes/no). The
original 7-point severity scale was reduced to a 6-point scale by combining
the categories “oral copulation of adult” and “oral copulation of child,”
which resulted in the following scale: (1) less severe abuse, (2) fondling/
molestation, (3) masturbation, (4) digital penetration, (5) oral copulation,
and (6) intercourse. Abuse duration is measured as a function of child age
(duration in days/age in days), so that comparisons could be made regard-
less of child age (e.g., a 3-year-old cannot experience a 5-year abuse).
Additional family problems. The presence of substance (alcohol or
drugs) abuse problems or dependence in the primary or secondary care-
giver and the presence of mental health problems in the primary caregiver
were reported by caregivers and child welfare workers. If either the worker
or the caregiver acknowledged these issues, the dichotomous variable was
coded “yes.” Child welfare workers also reported whether or not the family
had prior maltreatment reports and low social support at the time of the
investigation.
Caregivers reported their level of depression and substance use in the 12
months prior to and including the investigation using the Composite
International Diagnostic Interview—Short Form (CIDI–SF), a measure
with classification accuracy of 93% to 99% in relation to psychiatric
standards (Kessler, Andrews, Mroczek, Ustun, & Wittchen, 1998).
Caregivers reported interpersonal violence experiences in the 12 months
prior to and including the investigation using The Conflict Tactics Scale
(CTS1: Straus, 1979). Minor and severe violence subscales were collapsed
into one, DV, and combined with child welfare workers’ reports to create
the dichotomous measure of DV.
Child Behavior Checklist (CBCL; Achenbach, 1993). Internalizing and
externalizing behavior problems were assessed using the CBCL, a well-
validated measure indicative of psychopathology risk (Achenbach, 1993).
T scores at or above 65 are considered clinical-level problems, and scores
of 60 65 are considered borderline level. The CBCL has high interrater
reliability (.96), acceptable construct validity (.59 to .88), and high internal
consistency (.90 to .91) in the NSCAW sample (USDHHS, 2004).
Trauma Symptom Checklist for Children (TSCC; Briere, 1996). The
posttraumatic stress (PTS) subscale of the TSCC was administered to
children ages 8 years and older. The TSCC evaluates psychological symp-
tomatology in children who have experienced traumatic events (Briere,
1996). T scores at or above 65 are considered clinical level and scores
between 60 and 65 are borderline level. The PTS scale has high internal
consistency (␣⫽.85 to .87 across standard and maltreatment samples), and
the full TSCC shows construct validity in at least seven studies with regard
to traumatic impact (Briere, 1996).
Children’s Depression Inventory (CDI; Kovacs, 1992). The CDI is a
child-reported measure of depressive symptoms. Children are considered
depressed if they fall at or above the 91st percentile for their age and
gender group. In the NSCAW sample, internal consistency of the CIDI-SF
is acceptable, averaging .81 for 7- to 12-year-olds and .87 for 13-to
15-year-olds (USDHHS, 2004).
Services and case characteristics. Child welfare workers reported
whether any services were provided to the family, arranged, or referred
following the investigation (hereinafter termed provided or referred ser-
vices, in the interest of simplicity). Follow-up questions inquired about
particular services, including child mental health counseling and treatment.
Workers also reported the investigation status (substantiated, indicated, or
neither), whether or not the study child had serious emotional or behavioral
problems, and their living situation postinvestigation (in home or out of
home).
Data Analysis
Factor mixture modeling, a form of latent class analysis (LCA), and
regression analysis were the primary analytic methods. Factor mixture
modeling, using weights and adjusting for complex sampling, was con-
ducted using MPLUS statistical software, Version 3.11, and linear and
logistic regression analyses were conducted using SUDAAN software
(Research Triangle Institute, 2001). All analyses used the sampling weights
and clustering estimates unless otherwise noted.
Analyses were conducted in four steps. First, children were stratified
into three age groups: 3–7, 8 –11, and 12–14, with approximately one third
of the total sample falling into each group. Children in the three age groups
were compared with regard to abuse characteristics, presence of other
family problems, and psychological scores using chi-square statistics and t
tests.
Next, we conducted factor mixture modeling within child age groups.
Two statistical indicators of model fit (Bayesian Information Criterion
[BIC] and entropy) and three substantive indicators of model fit were used
in this study. Substantive value of the models were evaluated accordingly:
(a) the amount of new information gained with increased numbers of
classes, (b) statistical strength of the model relative to the information
gained, and (c) the estimated sample size of the additional class. Class
counts that were estimated to include fewer than 5% of the total children
in their respective age group (unweighted) were considered too small to be
stable or useful relative to the purposes of the study. After choosing the
470
MCCRAE, CHAPMAN, AND CHRIST
Page 3
best factor mixture models, we created substantive class labels, and the
demographic composition of each class was viewed.
The third analytic step was to conduct linear regression analysis mod-
eling the relationship between children’s latent class membership and their
psychological scores. Each outcome (internalizing and externalizing be-
havior, depression, and PTS) was modeled within each age group, resulting
in 10 models total (depression and PTS scores were not applicable to the
youngest children). Independent variables included four demographics
(child age, gender, race/ethnicity, and urbanicity), two case characteristics
(child’s living situation and substantiation status), the respective class
indicator variables, and a variable to control for time between the mal-
treatment report and the research interview. The latter was included to
account for psychological scores that may be attributable to children’s
proximity to the maltreatment or its discovery.
The final step was to conduct bivariate and multivariate analysis of the
mental health service response. Independent variables were those used in
the linear regression modeling (discussed earlier) and three additional
variables: mental health problems reported by child welfare workers, any
borderline- or clinical-level score, and latent class grouping indicative of
the most symptoms (classes with high symptom levels were collapsed and
compared with classes showing low levels of symptoms).
Results
Results of descriptive and multivariate analysis regarding vari-
ables used in the factor mixture modeling and children’s psycho-
logical scores are presented first, followed by results of bivariate
and multivariate analysis of services. Children in the study were
predominantly female (74%) and were White (51%), Black (23%),
Hispanic/Latino (18%), or other race/ethnicity (8%). Mean age of
children was 9 years, with 42% ages 3–7 years (unweighted n
196), 29% 8–11 years (n 177), and 29% 12–16 years (n 180).
Average time between the index maltreatment report and the
research interview was about 5– 6 months.
Bivariate Analysis of Variables Used in Factor Mixture
Modeling
Table 1 presents descriptive statistics for dependent variables
used in the factor mixture modeling (abuse and family problems)
and psychological outcomes including chi-square tests for age
group differences. Mean abuse severity for the population was 3.3
on the 6-point severity scale, a figure that was not significantly
different by child age. Abuse duration averaged 9 months, and the
youngest children were investigated for significantly shorter dura-
tion abuse, compared with the middle and oldest children (58 days,
457 days, and 374 days, respectively). Large standard errors were
generated with these figures, indicating substantial within-age
variance with regard to duration (e.g., SEs 197 and 95 among
the middle and oldest children, respectively).
Over 60% of children were investigated for abuse involving a
nonparental perpetrator: Thirty-three percent of children were in-
vestigated for relative-perpetrated abuse, and 34% of children were
investigated for abuse involving an unrelated person. The youngest
and middle children were more frequently investigated for relative
abuse, compared with the oldest children (40%, 43%, and 14%,
respectively). One quarter of children (26%) were investigated for
multitype maltreatment, with higher rates reported among 12- to
14-year-olds (41%), compared with 3- to 7-year-olds (11%).
Prior reports were common (56%), with nearly three quarters of
8- to 11-year-olds (71%) and over one half of 12- to 14-year-olds
(61%) having prior reports. Many 3- to 7-year-olds also had prior
Table 1
Children’s Psychological Scores, Abuse, and Family Characteristics by Child Age
Characteristic
Child Age (unweighted N)
Total
(N 553)
3–7
(n 196)
8–11
(n 177)
12–14
(n 180)
M (SE) M (SE) M (SE) M (SE)
Mean severity
a
3.2 (0.2) 3.0 (0.3) 3.7 (0.4) 3.3 (0.2)
Mean duration (days)
a
58 (19.1)
*
457 (197.0) 374 (95.0) 280 (62.6)
Perpetrator
Parent/figure 23 (4.2) 32 (7.9) 46 (11.4) 33 (4.8)
Other relative 40 (6.8)
*
43 (8.5)
*
14 (4.8) 33 (5.0)
All others 37 (5.7) 25 (9.6) 40 (11.1) 34 (6.8)
Multi type maltreatment 11 (3.4) 31 (8.9) 41 (11.4)
*
26 (4.5)
Prior maltreatment reports 42 (9.7) 71 (6.0) 61 (10.7) 56 (4.6)
Substance abuse 16 (4.9) 16 (5.5) 8 (2.5) 14 (2.9)
Mental illness 30 (5.4) 24 (5.6) 36 (10.6) 30 (3.3)
Domestic violence 45 (8.2) 44 (10.6) 38 (10.5) 43 (5.3)
Low social support 15 (4.9) 32 (9.3) 33 (10.8) 25 (4.8)
Borderline or clinical scores
Internalizing behavior 21 (6.7) 32 (9.8) 43 (11.3) 31 (5.1)
Externalizing behavior 34 (6.8) 42 (9.0) 67 (9.9) 46 (5.2)
Depression 12 (3.5) 23 (10.5) 20 (5.9)
b
Posttraumatic stress 24 (9.5) 14 (6.7) 19 (7.7)
1 borderline/clinical score 21 (6.6) 38 (9.0) 45 (11.0) 33 (4.4)
Any borderline/clinical score 40 (7.5) 62 (8.6) 80 (6.6)
*
58 (4.8)
a
Abuse severity scale ranges from 1– 6, with less severe indicated by lower numbers. Abuse duration reported
here is missing for 132 children.
b
Includes 7-year-old children.
*
p .05. All analyses are on weighted data.
471
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 4
reports (41%). Rates of caregiver substance abuse, MI, and DV
were 14%, 30%, and 43%, respectively. Low social support was
reported with regard to 25% of families.
Roughly one third to one half of children scored in the border-
line or clinical range of behavior problems (31% and 46% for
internalizing and externalizing behavior problems, respectively).
These figures are comparable with baseline rates among the total
NSCAW population wherein 24% to 36% of children, depending
on source, exhibited borderline or clinical-level internalizing
problems and 33% to 44% exhibited externalizing problems
(USDHHS, 2004). Mean internalizing score for children in the
present population was 54, and mean externalizing score was 57.
One fifth of children exhibited clinical-level depression (20%)
and PTS (19%). Rates of PTS varied significantly by child age,
with more 8- to 11-year-old children exhibiting PTS symptoms
(24%), compared with 12- to 14-year-old children (14%). Rates of
depression and PTS were higher among children in the present
study, compared with the larger NSCAW group, wherein baseline
rates of clinical-level depression and PTS were 15% and 12%,
respectively (Kolko et al., 2005; USDHHS, 2004). Mean depres-
sion and PTS scores were each 52 in this sample.
The majority of children had at least one borderline- or clinical-
level score (58%), and significantly more 12- to 14-year-old chil-
dren (80%) than 3- to 7-year-old children (40%) exhibited at least
one problem symptom. One third of children had comorbid-
problem-level symptoms (33%).
Results of Factor Mixture Modeling
Factor mixture modeling yielded two 5-class models (young-
est and oldest children) and one 4-class model (middle chil-
dren). As presented in Table 2, model fit for the youngest
children clearly improved from the 1-class model to the 2-, 3-,
and 4-class models, as shown by substantial drops in the BIC
with each additional class (values decreased by 143.90, 77.25,
and 62.22 units, respectively). The decrease in the BIC was
much less, however, in the models with 5 (24.19) and 6 (28.68)
classes, indicating that model parsimony improved negligibly
with these models.
Entropy, a measure of the extent to which the model distin-
guishes the classes, was acceptable regardless of the number of
classes, with values ranging from .93 (4- and 5-class models) to
.95 (6-class model). The 6-class model generated a class that
was small (11 children), below the established criterion of
having at least 5% of the children represented in a class.
Because substantively meaningful information was gained with
the 4-class model over the 3-class model (a class distinguished
by domestic violence), and again with the 5-class model (a class
with moderate severity abuse and caregiver MI), the 5-class
model was determined to be the best fit to the data for the
youngest children.
Data for children ages 8 –11 were well replicated by a 4-class
model. The BIC showed substantial decreases as classes were
added up to the 4-class model, with differences of 156.32 (1- to
2-class models), 104.44 (2- to 3-class models), and 72.48 (3- to
4-class models) units. However, the 5-class model showed a very
small drop in the BIC of just 4.38 units, indicating that the 4-class
model was adequately parsimonious. Entropy was the same in both
the 4- and 5-class models (.97). When 6 classes were imposed on
the data, the BIC further declined (23.37 units), but model entropy
was lower (.95), and the smallest class was estimated with 16
members. The 4-class model was determined to be the best fit to
the data for the children ages 8 –11.
Table 2
Fit Indices for the Factor Mixture Models
Number of classes BIC
Difference
in BIC Entropy
Largest/smallest class
size unweighted n
Ages 3–7
1 2233.80 196/0
2 2089.90 143.90 .94 178/18
3 2012.64 77.25 .94 110/28
4 1950.43 62.22 .93 68/16
5 1926.24 24.19 .93 56/24
6 1897.56 28.68 .95 55/11
Ages 8–11
1 2351.64 177/0
2 2195.32 156.32 .91 133/44
3 2090.88 104.44 .94 99/19
4 2018.40 72.48 .97 97/18
5 2014.02 4.38 .97 68/22
6 1990.65 23.37 .95 51/16
Ages 12–14
a
1 2183.47 180/0
2 2049.53 133.94 .92 93/87
4 1922.52 127.01 .95 94/16
5 1883.53 38.99 .97 80/14
6 1828.16 55.36 .97 57/16
Note. All analyses are on weighted data. BIC Bayesian Information Criterion.
a
The 3-class model for children in this age-group generated a class with 1 member, inhibiting the ability to
produce parameter estimates for this class and to conduct tests of model comparisons.
472
MCCRAE, CHAPMAN, AND CHRIST
Page 5
Some difficulty was encountered in the mixture modeling for
the oldest children. Namely, the 3-class solution yielded a class
with one member, limiting the ability to generate the parameters
for this class (no variance) and to conduct subsequent model
comparisons. As such, model fit for the surrounding models was
analyzed, with the conclusion that the 3-class model was inade-
quate, given the one-member class. The BIC value for the 4-class
solution decreased by 127.01 units from the 2-class model, indi-
cating better model fit in the 4-class model. Class distinction was
improved in the 4-class model (.95) over the 2-class model (.92).
The 5-class and 6-class models showed further drops in the BIC
(38.99 and 55.36 units, respectively) and constant entropy (.97).
The additional class in the 6-class model, however, was not sub-
stantially different than one generated in the 5-class model. The
5-class model was retained as the best fit to the data for children
ages 12–14.
Results of Factor Mixture Modeling: Class Distinctions
Results of the mixture modeling, in terms of the variables that
distinguish the classes, are presented in Tables 3 and 4. Table 3
figures are the proportion of class members who endorsed a
particular item for the dichotomous variables and the estimated
means generated by the latent class for the continuous variables.
Table 3 also presents the demographic composition of the classes
generated by a frequency distribution after the classes were esti-
mated. Table 4 presents the results substantively. Labels were
assigned to the classes on the basis of their most distinguishing
characteristics. Classes within each age group were ordered from
the largest to smallest class size and numbered starting at Class 1
(normative). The normative class was the most common and,
regardless of child age, was characterized by the least severe
sexual abuse of the shortest duration and comparatively low en-
dorsement of other family problems.
Children ages 3–7 years. The normative class comprised 29%
of this population and was characterized by less severe abuse (␮⫽
1.66) that allegedly occurred one time by a nonparental relative or
unrelated person. No other family problems characterized this
class, and the group was gender and racially diverse.
Class 2 comprised another 26% of children and was distin-
guished by DV (89% endorsement) and caregiver MI (60%). Class
2 members were investigated for less severe (␮⫽1.99), shorter
duration abuse (␮⫽.01) and were primarily White (48%) and
Black (37%) girls (83%). Children in Classes 3, 4, and 5 were
investigated for more severe CSA, on average, compared with
children in the first 2 classes. Class 3 members (16%) were
investigated for the longest duration abuse (␮⫽.30), and were
similar to Class 4 members (16%) in terms of abuse severity (s
5.02 and 5.57, respectively). Class 4 members were investigated
for shorter duration abuse, often had prior reports (84%), and were
mostly White (64%) girls (88%). Class 5 members (13%) were
mostly boys (66%) investigated for moderate severity (␮⫽3.85),
shorter duration abuse (␮⫽.01) by a nonparental relative (.75).
Classes were labeled accordingly: least severe, DV/MI, severe
chronic, severe moderate, and boys.
Children ages 8 –11. The normative class comprised over one
half of 8- to 11-year-olds (54%), and these children were investi-
gated for less severe (␮⫽1.80), shorter duration abuse (␮⫽.02)
involving a variety of perpetrators, including parent or parent
figures (35%), other relatives (24%), and unrelated persons (41%).
Over one half of Class 1 members had caregivers with DV expe-
riences (54%).
Class 2 members (22%) were investigated for more severe abuse
(␮⫽5.17) over a moderate duration (␮⫽.09) by a nonparental
relative (90%). Nearly three quarters of Class 2 members have had
prior reports (70%), and 44% were investigated for multitype
maltreatment. Class 3 members (14%) were also investigated for
more severe abuse (␮⫽5.69), but over a longer duration (␮⫽
.24) by a parent or parent figure (96%). Class 3 members highly
endorsed family problem items, including low social support
(78%), prior reports (75%), substance abuse (70%), and DV
(60%).
The last class among 8- to 11-year-olds (Class 4; 10%) was
distinguished by less severe abuse (␮⫽2.01) over a long duration
(␮⫽.474) by a nonparental relative (100%). Class 4 members
were nearly all White (97%) girls (92%) with prior maltreatment
reports (100%) and family low social support (89%). Classes were
labeled accordingly: least severe, severe moderate, severe all, and
chronic relative.
Children ages 12–14. The normative class among 12- to 14-
year-olds comprised 45% of the population and was characterized
by less severe abuse (␮⫽1.91) over a shorter duration (␮⫽.03)
by a parent or parent figure (60%) or unrelated person (23%). Most
members were girls (78%) and were White (55%) or Hispanic or
from other races/ethnicities (36%). Caregiver DV was frequently
endorsed (39%).
Class 2 members (26%) were investigated for higher severity
abuse (␮⫽5.70) over a longer duration (␮⫽.14) by a variety of
perpetrators, including parent or parent figures (42%), relatives
(23%) and unrelated persons (35%). Class 2 members were ra-
cially diverse, and 43% had a caregiver with MI. Class 3 members
(12%) were also investigated for more severe CSA (␮⫽5.96), but
which occurred over a shorter duration (␮⫽.05), by an unrelated
person (99%). Class 3 members were nearly all Black (91%) girls
(100%) with high endorsement of family problems, including prior
reports (100%), multitype maltreatment (97%), low social support
(96%), substance abuse (100%), and DV (100%).
The final two latent classes among 12- to 14-year-olds com-
prised 17% of the population. Class 4 members (9%) were again
mostly Black (79%) girls (93%) investigated for moderate severity
abuse (␮⫽3.19) over a shorter duration (␮⫽.02) by an unrelated
person (86%). Nearly all Class 4 members had prior reports (90%)
and caregivers with MI (85%). Class 5 members (8%) were in-
vestigated for moderately severe abuse (␮⫽3.19) over a long
duration (␮⫽.10) by a parent or parent figure (100%). Class 5
members were nearly all White boys (93%) and highly endorsed
multitype maltreatment (94%), prior reports (91%), low social
support (99%) and caregiver MI (90%). Classes were labeled
accordingly: least severe, severe chronic, severe all, MI/priors, and
boys.
Summary of class distinctions. Regardless of age, the largest
groups of children were investigated for less severe abuse that
reportedly occurred over a short duration or on one occasion.
These classes comprised about one half of children in each age
group and typically did not show high endorsement of other family
problems (except that DV was prevalent among younger children).
Two classes in each age group were distinguished by more severe
abuse (involving penetration) that allegedly occurred over either a
473
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 6
Table 3
Results of Factor Mixture Modeling: Variables that Distinguish the Classes
Class number
a
(unweighted n)
Estimated
percent of
children
Severity
(M)
Duration
b
(M)
Proportion of members endorsing the item
Demographic
Composition (%)
a
Parent/parent
figure
perpetrator
Other
relative
perpetrator
Multi-type
maltreatment
Prior
reports
Low social
support
Substance
abuse MI DV Male Black Hispanic
Ages 3–7 (n 196)
1 (normative) 29 1.66 .004 .06 .59 .02 .29 0 .02 0 0 30 26 23
2 26 1.99 .010 .47 .05 .18 .34 .17 .37 .60 .89 17 37 15
3 16 5.02 .298 .39 .35 .09 .16 .22 .09 .14 .57 6 27 27
4 16 5.57 .021 .17 .38 .26 .84 .42 .22 .18 .57 12 7 29
5 13 3.85 .006 0 .75 .00 .69 0 0 .70 .24 66 7 33
Ages 8–11 (n 177)
1 (normative) 54 1.80 .022 .35 .24 .35 .65 .15 .07 .25 .54 40 25 38
2 22 5.17 .086 0 .90 .44 .70 .18 .12 .28 .25 48 15 36
3 14 5.69 .236 .96 0 .22 .75 .78 .70 .31 .60 18 25 15
4 10 2.01 .474 0 1.0 .01 1.0 .89 0 .01 .01 8 0 3
Ages 12–14 (n 180)
1 (normative) 45 1.91 .030 .60 .17 .34 .47 .13 .05 .21 .39 22 9 36
2 26 5.70 .138 .42 .23 .25 .47 .30 .13 .43 .35 8 13 27
3 12 5.96 .051 .01 0 .97 1.0 .96 1.0 0 1.0 0 91 0
4 9 3.19 .023 0 .14 .08 .90 0 .05 .85 0 7 79 0
5 8 4.02 .099 1.0 0 .94 .91 .99 .04 .90 .07 93 4 3
Note. All analyses are on weighted data. MI mental illness; DV domestic violence.
a
Based on most likely class membership, generated by a frequency distribution after the classes were formed.
b
Duration in days/child age in days.
474
MCCRAE, CHAPMAN, AND CHRIST
Page 7
shorter or longer duration. Neither group (severe short-duration
abuse or severe long-duration abuse) showed consistent patterns
with regard to family problems, multitype maltreatment, or
perpetrator. Classes with these higher severity acts were prev-
alent, comprising about one third of children in each age group.
The final abuse-related pattern was that classes generated in the
models for older children tended to show perpetrator distinc-
tion, whereas classes in the models for the youngest children
were not.
Also evident across the mixture models were several classes
distinguished by caregiver MI and DV (not necessarily together).
Classes distinguished by substance abuse, however, were nearly
always accompanied by DV. Prior maltreatment reports character-
ized nearly all children (and classes) in the middle and oldest age
groups.
Demographic composition of the classes showed that two com-
prised mostly boys—a group of 3- to 7-year-old boys and a group
of 12 to 14-year-old boys. Both were characterized by moderate
severity, shorter duration abuse and high endorsement of caregiver
MI. Most classes were racially diverse, except that two classes of
12- to 14-year-olds were mostly Black girls.
Results of Linear Regression Modeling Children’s
Psychological Symptoms
Table 5 presents the results of linear regression analysis mod-
eling children’s psychological scores. Models presented for each
age group exclude the normative (largest) class, such that the
parameters represent the relationship between the respective class
and the normative class. Results of models that tested all of the
class comparisons are not presented in Table 5, but significant
findings will be discussed. Children were assigned to their most
likely class membership on the basis of their latent class proba-
bilities (0% to 100%). Very few (1%) midrange probability
values (.20 –.80) were observed.
Children ages 3–7. Class 2 (DV/MI) and Class 5 members
(boys) scored significantly higher than other children in terms of
behavior problems. Children in the DV/MI group showed increased
internalizing and externalizing scores of 6.34 and 6.68 points, respec-
tively, over children in the normative class (least severe). Class 2
internalizing scores were also significantly higher than Class 4
(severe moderate) scores. Class 5 (boys) members scored sig-
nificantly higher than normative and Class 4 members in terms
of externalizing behavior problems. Children with substantiated
Table 4
Conceptual View of Class Characteristics
Class Label Description
Ages 3–7
1 (Normative) Single
incident
Gender and racially diverse group investigated for less severe, typically single incident abuse by a non-parental
relative or unrelated perpetrator. Very low incidence of other reported family problems.
2 DV/MI White and Black girls investigated for less severe abuse over a short duration by a parent or parent figure or
unrelated person. Nearly all have caregivers with DV experiences and over 50% have caregivers with MI.
3 Severe chronic Racially diverse girls investigated for more severe abuse over a long duration. Perpetrator varies by group
member. Over 50% have caregivers with DV experiences.
4 Severe moderate Mostly White girls investigated for more severe abuse over a moderate duration by a non-parental relative or
unrelated person. Nearly all have prior reports, and DV and low social support characterize some families.
5 Boys Mostly White and Hispanic/other boys investigated for moderately severe abuse over a short duration by a
non-parental relative. Roughly two-thirds have a caregiver with MI and prior reports.
Ages 8–11
1 (Normative) Single
incident
Gender and racially diverse group investigated for less severe, typically single incident abuse. Perpetrator
varies by group member. Over 50% have caregivers with DV experiences.
2 Severe moderate Gender and racially diverse group investigated for more severe abuse over a moderate duration by a non-
parental relative. Multi-type maltreatment is common (44%), and 70% have prior reports.
3 Severe all Mostly girls with some racial diversity investigated for more severe abuse over a long duration by a parent or
parent figure. Multiple other family problems evident. Highest rates of substance abuse and DV.
4 Chronic relative White girls investigated for less severe abuse over a long duration by a non-parental relative. All have prior
reports and most live in families with low social support. Very low incidence of caregiver substance abuse,
MI and DV.
Ages 12–14
1 (Normative) Least
severe
White and Hispanic/other girls investigated for less severe abuse over a short duration by a parent or parent
figure or unrelated perpetrator. Over two-thirds (39%) have caregivers with DV experiences.
2 Severe chronic Girls with some racial diversity investigated for more severe abuse over a long duration. Perpetrator varies by
group member. Nearly one-half (43%) have a caregiver with MI.
3 Severe all Black girls investigated for more severe abuse over a moderate duration by an unrelated perpetrator. Multiple
other family problems (except caregiver MI) characterize these families.
4 MI/priors Black girls investigated for moderately severe abuse over a short duration by an unrelated perpetrator. Nearly
all have a caregiver with MI and prior reports.
5 Boys White boys investigated for moderately severe abuse over a long duration by a parent or parent figure. Nearly
all live in families with low social support, multi-type maltreatment, prior reports, and caregiver MI.
Note. DV domestic violence; MI mental illness.
475
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 8
or indicated abuse had significantly lower internalizing scores,
compared with children whose abuse was unsubstantiated.
Children ages 8 –11. Class 4 (chronic relative) members
scored significantly higher than children in all other latent classes,
with regard to externalizing behavior problems and PTS, and
higher than Class 2 (severe moderate) members in terms of de-
pressive symptoms. Class 3 members (severe all) had significantly
higher internalizing scores of nearly 5 points, compared with
normative-class children and significantly higher depression
scores (13.81 points) than Class 2 members. Finally, internalizing
scores were over 7 points higher among Black children, compared
with other children.
In summary, children ages 8 –11 in Class 4 (chronic relative)
showed the highest externalizing and PTS symptoms, whereas
internalizing and depressive symptoms were notable among Class
3 (severe all) and Class 4 members.
Children ages 12–14. Children in Class 5 (boys) scored sig-
nificantly higher than children in all other classes with regard to
externalizing behavior problems and depressive symptoms and
higher than normative and Class 2 members (severe chronic) in
terms of internalizing problems and PTS. Class 5 members also
scored higher than Class 4 members in terms of PTS. Class 4
members (MI/priors) scored significantly higher than Class 3
members, with regard to externalizing and depression scores, and
higher than Class 2 members in terms of internalizing behavior
problems. Normative and Class 2 members appeared to show the
least symptoms overall, although Class 2 members showed greater
externalizing and PTS symptoms than normative class members.
Children living in out-of-home care after the investigation
scored over 9 points higher than children who remained at home in
terms of internalizing behavior problems. Externalizing behavior
scores were increased by over 5 points in association with sub-
Table 5
Linear Regression Modeling Children’s Psychological Scores
Independent variable
Internalizing
(SE)
Externalizing
(SE)
Depression
(SE)
PTS
(SE)
Ages 3–7 (n 196)
Time to interview in months 0.52 (0.37) 0.46 (0.32)
Child age in months 0.06 (0.09) 0.07 (0.10)
Child gender: Male 3.40 (2.6) 3.39 (2.6)
Child race: Black 4.61 (2.6) 4.54 (3.2)
Child race: Hispanic/other 3.37 (2.8) 0.84 (3.3)
Out-of-home placement: Yes 5.39 (3.8) 4.72 (2.6)
Urbanicity: Nonurban 1.05 (2.9) 0.09 (2.7)
Substantiated or indicated: Yes 4.54 (1.6)
*
4.16 (2.2)
Class 2 (DV/MI)
a
6.34 (2.4)
*
6.68 (3.4)
*
——
Class 3 (severe chronic) 2.51 (2.9) 3.99 (6.0)
Class 4 (severe moderate) 1.55 (3.6) 0.75 (3.4)
Class 5 (boys) 6.54 (4.6) 7.87 (3.8)
*
——
Model R
2
0.28 0.22
Ages 8–11 (n 177)
Time to interview in months 0.00 (0.38) 0.94 (0.67) 0.98 (0.79) 0.93 (0.50)
Child age in months 0.06 (0.08) 0.01 (0.10) 0.00 (0.10) 0.19 (0.10)
Child gender: Male 4.98 (3.5) 0.75 (3.0) 0.93 (2.3) 3.88 (4.0)
Child race: Black 7.49 (3.2)
*
7.53 (3.9) 0.72 (2.7) 3.91 (3.8)
Child race: Hispanic/other 2.25 (2.7) 3.34 (2.5) 2.31 (2.8) 2.90 (2.3)
Out-of-home placement: Yes 3.80 (2.2) 3.42 (3.7) 0.61 (2.9) 1.25 (2.6)
Urbanicity: Nonurban 5.54 (3.0) 2.88 (2.8) 1.32 (2.6) 1.57 (3.5)
Substantiated or indicated: Yes 0.75 (2.6) 4.33 (2.8) 2.71 (3.5) 0.16 (3.7)
Class 2 (severe moderate)
a
2.42 (2.9) 2.44 (2.6) 4.42 (3.2) 1.31 (3.4)
Class 3 (severe all) 5.40 (2.4)
*
0.93 (4.3) 9.38 (6.2) 1.17 (4.2)
Class 4 (chronic relative) 3.30 (2.5) 10.01 (2.7)
*
7.27 (4.5) 29.82 (5.4)
*
Model R
2
0.21 0.31 0.28 0.50
Ages 12–14 (n 180)
Time to interview in months 0.82 (0.50) 1.37 (0.50)
*
1.03 (0.50)
*
0.13 (0.48)
Child age in months 0.08 (0.12) 0.22 (0.12) 0.20 (0.09)
*
0.06 (0.08)
Child gender: Male 0.14 (4.2) 3.01 (4.2) 12.76 (3.4)
*
0.34 (2.5)
Child race: Black 2.76 (2.2) 4.27 (2.3) 4.01 (3.3) 1.08 (2.3)
Child race: Hispanic/other 5.43 (3.5) 4.74 (3.9) 10.05 (4.5)
*
6.94 (2.6)
*
Out-of-home placement: Yes 9.78 (3.6)
*
2.84 (2.9) 3.81 (2.7) 2.32 (2.3)
Urbanicity: Nonurban 3.96 (2.8) 3.45 (3.2) 4.67 (2.9) 1.94 (2.8)
Substantiated or indicated: Yes 2.39 (2.4) 5.84 (2.4)
*
4.36 (2.7) 0.07 (1.7)
Class 2 (severe chronic)
a
3.40 (2.6) 7.29 (2.6)
*
4.01 (2.6) 6.54 (2.2)
*
Class 3 (severe all) 9.74 (2.9)
*
1.39 (3.5) 8.51 (4.7) 10.44 (3.9)
*
Class 4 (MI/priors) 9.47 (2.9)
*
10.08 (3.9)
*
1.20 (5.7) 6.67 (3.9)
Class 5 (boys) 14.38 (4.2)
*
24.41 (4.5)
*
17.42 (3.0)
*
16.00 (2.8)
*
Model R
2
0.48 0.55 0.42 0.37
Note. DV domestic violence; MI mental illness; PTS posttraumatic stress.
a
Based on most likely class membership. Class 1 (normative) is the referent group.
*
p .05. All analyses are on weighted data.
476
MCCRAE, CHAPMAN, AND CHRIST
Page 9
stantiated or indicated abuse. More depressive and PTS symptoms
were reported by Hispanic children and children of other races/
ethnicities, compared with White and Black children. Boys
showed significantly fewer depressive symptoms than girls.
Results of Services Analyses
Table 6 presents the results of bivariate and multivariate anal-
yses of children’s referral to, or receipt of, mental health services.
Overall, 58% of children scored in the borderline- or clinical-level
range of at least one symptom, and 43% were referred to or
provided mental health services. Just 26% of child welfare workers
reported that the referent child had serious emotional or behavior
problems. Service referral or receipt rates were 36%, 51%, and
46% among children ages 3–7, 8 –11, and 12–14, respectively.
Rates of mental health service referral or receipt were not signif-
icant in bivariate analyses (chi-square statistics) according to chil-
dren’s latent class grouping, exhibited symptoms, or worker-
reported mental health problems.
Most abuse was unsubstantiated or not indicated (61%). In
bivariate analyses, substantiated or indicated abuse was uniformly
related to mental health service referral; 72% of children with
substantiated/indicated abuse were referred to or received mental
health services, compared with 24% among children whose abuse
was unsubstantiated or not indicated. Among children ages 3–7,
58% of children with substantiated or indicated abuse were re-
ferred to or provided mental health services, in comparison with
20% of children whose abuse was unsubstantiated or not indicated.
Service proportions were 82% and 81%, respectively, among
children ages 8–11 and 12–14 with substantiated or indicated
abuse, compared with 33% and 22%, respectively, of unsubstan-
tiated or not indicated cases. Most children remained at home after
the investigation (90%), and mental health service referral rates
were higher in association with out-of-home placement (72%),
compared with remaining at home (40%).
Results of the three logistic regression analyses modeling men-
tal health service referral or receipt are also presented in Table 6.
Because of problems with model instability due to some classes
with few members (unweighted), it was necessary to combine
some of the latent classes in these models. The decision was made
to combine classes on the basis of symptomatology (groups of
children with the highest symptom levels were combined). Among
children ages 3–7, Classes 2 and 5 were combined and compared
with Classes 1, 3, and 4. Among children ages 8–11, Classes 3 and
4 were combined and compared with Classes 1 and 2. Classes 2 to
5 were combined and compared with the normative class among
children ages 12–14.
Results show that, regardless of child age, substantiation status
of the CSA explained much of the variance in mental health
service referral rates. Odds of mental health service referral or
receipt were increased by 4.73 among 3- to 7-year-olds with
substantiated or indicated abuse over children whose abuse was
not substantiated, and the estimates were higher among 8- to
11-year-olds and 12- to 14-year-olds (OR 12.73 and 28.05,
respectively).
Two other variables predicted mental health service referral
among the youngest children. Children whose child welfare work-
ers reported a mental health problem were over four times more
likely to have been referred to or received mental health services,
compared with children whose child welfare workers did not
report a problem (OR 5.67). Odds of mental health service
referral were significantly reduced among children living in non-
urban communities, compared with children living in urban com-
munities (OR 0.13).
Among 8- to 11-year-old children, apart from substantiation
status, mental health service referral or receipt was more likely in
association with CWW-reported child mental health problems
(OR 2.75). Among 12- to 14-year-olds, children with at least
one borderline or clinical symptom had greater odds of mental
Table 6
Results of Logistic Regression Analyses Modeling Mental Health Service Referral or Receipt
Case characteristics
Post-investigation mental health service referral or receipt
Ages 3–7 (n 190) Ages 8–11 (n 168) Ages 12–14 (n 175)
% Yes
Odds
ratio 95% CI % Yes
Odds
ratio 95% CI % Yes
Odds
ratio 95% CI
Time to interview in months NA 1.08 .85, 1.36 NA .92 .72, 1.18 NA .86 .68, 1.09
Child age in months NA 1.03 .99, 1.07 NA 1.00 .96, 1.04 NA .98 .94, 1.03
Gender: Male 34 .84 .23, 3.09 42 1.08 .13, 8.69 22 .54 .13, 2.18
Race: Black 18 .24 .03, 2.32 66 2.48 .32, 19.19 10 .12
*
.03, .52
Race: Hispanic/other 60 2.39 .34, 5.62 34 .59 .20, 1.76 76 5.90 .85, 40.79
Out-of-home placement: Yes 80
*
4.35 .86, 21.94 88 5.78 .66, 50.95 54 2.36 .75, 7.47
Urbanicity: Nonurban 11 .13
*
.02, .93 47 .50 .13, 1.96 24 .20 .04, 1.09
CWW report MH problem: Yes 73 5.67
*
1.31, 24.56 63 2.75
*
1.03, 7.35 36 1.06 .32, 3.56
Substantiated or indicated: Yes 58
*
5.73
*
1.98, 16.55 82
*
12.73
*
2.76, 58.66 81
*
28.05
*
9.05, 86.94
Any borderline/clinical symptoms 39 1.39 .49, 3.98 49 .66 .09, 4.67 52 7.70
*
1.19, 50.04
Class(es) with high scores
a
33 1.37 .44, 4.26 83 1.25 .19, 8.46 36 .43 .11, 1.67
Total %, Model R
2
36 .36 51 .33 46 .54
Note. NA not applicable; CWW child welfare worker; MH mental health.
a
Corresponds with Classes 2 and 5 among 3–7-year-olds, Classes 3 and 4 among 8–11-year-olds, and Classes 2–5 among 12–14-year-olds. Comparison
group includes the remaining classes. Bivariate tests of services and class membership that included all the classes were not significant.
*
p .05 using bivariate (chi-square) and multivariate analyses. All analyses are on weighted data.
477
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 10
health service referral or receipt, compared with children whose
scores were all in the normal range (OR 7.70). Black children
were over 80% less likely to have been referred to or received
mental health services, compared with other children (OR 0.12).
Limitations
Nearly one quarter of the study children (24%, unweighted)
were missing data regarding the duration and frequency of the
abuse, and another 36 children had seemingly incongruent infor-
mation (e.g., duration of 6 months, but one-time frequency). The
frequency variable was dropped from the analyses, which may
have resulted in some inaccuracy about abuse intensity, although
this is a slight possibility because these variables are strongly
correlated in these (r .65) and other analyses (Ruggiero et al.,
2000). Children missing the duration variable were maintained in
the study with missing data generation techniques (maximum
likelihood estimation under missing at random conditions; see
Muthe´n & Muthe´n, 2000), but it is possible that, had the duration
estimates in missing cases come directly from child welfare work-
ers, the results would be slightly different.
Measurement effects from using different raters to assess psy-
chological symptoms may be present. Behavior problems were
reported by caregivers, and depression and PTS were reported by
children. A strength of these data, however, is that children are
able to “speak for themselves,” so although measurement effects
are possible, we believe that having children describe their own
symptoms whenever possible outweighs these concerns. Other
NSCAW findings show differences of 10% to 12% in rates of child
behavior problems by rater (USDHHS, 2004).
Finally, the findings only apply to children’s symptoms imme-
diate to an investigation for sexual abuse and cannot be interpreted
to mean that their abuse or other experiences caused their symp-
toms because of the cross-sectional nature of the data. However,
the study offers a solid view of children’s psychological needs at
intake to CWS because commonly cited psychological problems
following CSA (e.g., aggressive or sexualized behaviors, depres-
sion) are likely to have been captured.
Discussion
This study contributes to what we know about children’s psy-
chological service needs following sexual abuse in two important
ways: It provides information about patterns of sexual abuse and
other family problems that relate to different types of psycholog-
ical symptoms, and it links these profiles to service receipt and
other CWS patterns that contribute to children receiving mental
health services.
Most fundamentally, the study demonstrates that not all inves-
tigations of CSA are alike. Even within rather narrowly defined
age groups, children’s experiences vary greatly. Although most
children were investigated for lower severity CSA that reportedly
occurred one time, many other children reportedly experienced
abuse over the majority of their lifetime. Prior maltreatment re-
ports were common (56%), whereas a fair number of children were
referred for maltreatment for the first time in adolescence (39%).
Five distinct subgroups of 3- to 7-year-olds and adolescents were
identified, along with four subgroups of school-age children. It is
clear that the practice of classifying children according to the
predominant type of maltreatment, common in child welfare ser-
vices (USDHHS, 2002), likely diminishes the quality of our un-
derstanding of children’s service needs. The finding in this study
that many children have prior reports and lower severity CSA,
along with family problems such as domestic violence (43%) and
caregiver MI (30%) indicate that, as a group, sexually abused
children may have experiences leading up to the referral point that
are more similar than not to the experiences of other maltreated
children referred to CWS. Efforts to design CWS that are more
developmentally sensitive (Berrick et al., 1998) will also require
sensitivity to wide variability in children’s experiences within their
peer and “maltreatment type” groups.
Prior research suggests that the worst psychological outcomes
are to be expected among children who experience a longer dura-
tion of abuse, abuse by a parental perpetrator, and abuse that
involves more invasive sexual acts (Kendall-Tackett et al., 1993;
Trickett, Noll, Reiffman, & Putnam, 2001). This study demon-
strates that these characteristics alone may not be a reliable gauge
of children’s service needs. Certainly, severity of abuse was an
important distinguishing variable in the classes that resulted from
this analysis, yet, the role of the other characteristics of the
maltreatment situation are also highlighted. Young children in this
study (ages 3 to 7) who demonstrate the most behavioral symp-
toms are those whose caregivers have domestic violence experi-
ences and MI. Other 3- to 7-year-olds that would be expected to
show psychological difficulties do not—namely, children whose
sexual abuse was substantiated or indicated and children whose
abuse was reportedly severe and chronic. Substantiated or indi-
cated abuse in this age group was associated with a 4-point
decrease in internalizing symptoms, and severe chronic abuse did
not predict internalizing or externalizing symptoms. These find-
ings may reflect the ability of young children to recover from
sexual abuse when the abuse is stopped, caregivers are able to be
supportive, and other family problems are not present. Alterna-
tively, these findings could be testament of sleeper effects, wherein
these children will later develop problems, or indication that the
discovery of the abuse at this point may prevent children from
developing symptoms altogether. It is clear that the mix of domes-
tic violence, caregiver MI, and low severity sexual abuse distin-
guishes a group of young children for whom behavioral symptoms
are a concern.
Other findings indicate that the combination of sexual abuse and
other family problems is important to understanding children’s
psychological needs. School-age children (ages 8 to 11) with more
severe sexual abuse (e.g., incest) along with multiple other family
problems, including caregiver substance abuse and domestic vio-
lence, exhibit internalizing and depressive symptoms to a greater
degree than other children. Externalizing behavior problems and
PTS symptoms, in contrast, are elevated among 8- to 11-year-old
White girls who experience chronic CSA by a nonparental relative.
This group is also distinguished by having prior maltreatment
reports, suggesting that the abuse was ongoing despite repeated
referrals to CWS. The development of externalizing and stress-
related symptoms in these children may, in turn, be related not
only the abuse itself but also to feelings or reactions (e.g., help-
lessness) that children may have to the repeated failure of adults to
protect them from the abuse. “Acting-out” behaviors among these
children may be in direct response to the sexual abuse or, in some
manner, may represent attempts by these children to bring atten-
478
MCCRAE, CHAPMAN, AND CHRIST
Page 11
tion to their plight. However, this is speculation; the need for
additional inquiry is indicated by this finding.
If any group of CSA-exposed children ought to be uniformly
provided mental health services, adolescents, of whom 80% scored
in the borderline or clinical range on at least one psychological
measure, would appear to be the appropriate group. The highest
rates of symptoms (except depression) were observed among chil-
dren whose CSA experiences were severe (e.g., oral sex, inter-
course). Depressive symptoms were associated with gender (more
girls than boys) and were notably higher among Hispanic and/or
other racially identified children. One group of boys in this age
group, in particular, showed elevated symptoms across all mea-
sures—White boys investigated for chronic sexual abuse involving
a parent or parent figure. These children clearly experienced co-
morbid or multiple symptoms at the time of the investigation.
Adolescent girls who show increased externalizing behavior prob-
lems and PTS but do not appear to experience internalizing and
depressive symptoms typically have abuse experiences that are
severe and chronic and are also likely to have a primary caregiver
with MI. PTS symptoms are elevated among adolescent Black girls
whose sexual abuse likely constitutes rape by a person unrelated to
them. These girls also experience multiple family problems.
Despite identifying distinct subgroups of children who exhibit
psychological needs, findings also show that children’s class mem-
bership does not fully explain some symptoms. Black children
ages 8 to 11, for example, show elevated behavioral symptoms
regardless of their class membership. The finding that Hispanic
and other nonmajority adolescents have greater levels of depres-
sion and PTS, regardless of class membership, suggests there are
other factors that contribute to these patterns. These findings may
relate to cultural differences surrounding the experience or disclo-
sure of sexual maltreatment, particular cultural views regarding
sexual maltreatment that would heighten levels of shame in vic-
tims, or institutional discrimination or help-seeking preferences
among families that culminate in service disparities and unad-
dressed psychological problems over time. Further research is
needed to understand these differences.
It is also important to note that, although many children referred
for sexual abuse exhibit symptoms at the time of the investigation
(58%), their symptom rates do not appear notably different in
comparison with other children seen by CWS. Internalizing be-
havior problems, for example, were noted among 31% of the
children in the present study and 32% among 4- to 14-year-olds in
the larger NSCAW group (USDHHS, 2004). Rates of clinical-
level depression were 20% and 15%, respectively, among the
present and total NSCAW children, and investigators using the
NSCAW data found no relationship between clinical-level PTS
and investigations of sexual abuse, physical abuse, and neglect
(Kolko et al., 2005). Certainly, many children in the present study
may develop problems, and their rates of psychopathology may
eventually exceed rates of other maltreated children—Kendall-
Tackett and colleagues (1993) estimated that between 10% and
20% of sexually abused children may develop problems a year or
more after the abuse— but at least initially, CSA-referred children
may not be substantially different in terms of their need for mental
health services, compared with other maltreated children.
Other findings from the NSCAW study show that, although
children’s clinical needs drive mental health services, as per be-
havioral indication using the CBCL (Achenbach, 1993), many
children who need services do not receive them (Burns et al., 2004;
Hurlburt et al., 2004; USDHHS, 2004). The present study suggests
that the opposite may often be true in cases of sexual abuse. Many
children who receive or are referred to services do not show an
immediate need. More children in the present study were referred
to or provided mental health services at the time of the investiga-
tion (43%) than children in the total NSCAW sample who received
such services in the 12 months before (16%) and after (28%)
intake to CWS (Burns et al., 2004; Hurlburt et al., 2004).
In addition, more children in the present study were referred to
or provided mental health services than their child welfare worker
reported as having emotional or behavioral problems (26%). This
suggests that many children are provided services for some reason
other than acute need. The possibilities include prevention of
symptoms or referrals that relate to worker or caregiver anxiety
about the sequelae of sexual abuse. This is not a surprising finding,
but it supports concerns raised by other investigators that more
ought to be learned about the utility of providing mental health
services to CSA-exposed children when circumstances or needs do
not dictate (Finkelhor & Berliner, 1995). Unquestionably, preven-
tive services are often warranted. However, even the most effica-
cious interventions for sexually abused children are designed to
address symptoms (Cohen & Mannarino, 1996; Saunders, Ber-
liner, & Hanson, 2001). This work supports the notion that CSA
often results in a mental health service response whether there is
clear evidence of need or not. For a system that is constantly
burdened by a lack of resources, appropriate targeting of services
would appear to be a critical goal. In turn, however, research is
needed to understand how and when to use preventative mental
health services with sexually abused children. These findings
suggest that other maltreated children may receive benefits from
similar services.
Results are clear, across age groups, that provision of mental
health services are driven by the substantiation status of the CSA
rather than the children’s exhibited needs or the totality of circum-
stances (e.g., abuse, prior reports, caregiver MI) that likely influ-
ence their psychopathology risk. Symptoms were predictive of
mental health services only among adolescents, and still, the
greatest likelihood of services among this group was in association
with substantiated or indicated abuse. Young children, in particu-
lar, with borderline or clinical-level symptoms were no more likely
to receive or be referred to mental health services than asymptom-
atic children. Service rates increased when children’s child welfare
workers reported problems, but very few young children (ages 3 to
7) were perceived by workers to have serious emotional or behav-
ioral problems (8%). The level of incongruence between young
children’s psychological needs reported by caregivers and worker
reports of problems strongly suggests that underidentification of
young children’s psychological needs is a primary concern, a
finding that may help explain other NSCAW findings that show
young children are unlikely to receive MHS (Burns et al., 2004).
In addition to the probability that children with unsubstantiated
sexual abuse and young children have unmet mental health service
needs after the investigation, other demographic characteristics
seem to be associated with unmet psychological needs, including
(a) children remaining at home, (b) children living in rural com-
munities, and (c) Black children. Two groups of adolescents are
seemingly underserved: White boys investigated for parent-
perpetrated severe abuse accompanied by multiple other family
479
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 12
problems, and Black girls with more severe abuse by an unrelated
person and multiple family problems. Service referral and receipt
rates were just 4% to 7%, respectively, among children in these
groups. Hispanic adolescents and children of other races/ethnici-
ties, in contrast, appear to be connected to services when they
exhibit symptoms.
Designing child welfare services that better protect the mental
health of sexually abused children will require interventions that
meet the individual needs of children and families. This study
shows that at least some children receive a mental health service
response based more on worker concerns when abuse is substan-
tiated rather than on specific assessment of children’s individual
needs. Factors that might account for this are agency resources that
are accessible in concert with substantiated abuse or family disin-
terest in receiving or following through with mental health services
when it is not clear that the abuse took place. Because there is “no
specific test that can convincingly conclude that sexual abuse has
not occurred” (Runyan, 1998, p. 494) and because there is evi-
dence that substantiation status is not a reliable gauge of maltreat-
ment risk, nonoccurrence, or family needs (Drake, 1995), services
that are guided by indicators other than child need are likely either
to underestimate the number of children who would benefit from
services or to misdirect resources.
Children’s psychological symptoms do show relation to patterns
of both their exposure to sexual abuse and other family problems,
supporting the provision of services that are abuse-informed but
not exclusively abuse-focused (Saunders et al., 2001). Moving
child welfare practice toward service provision that is based on a
tailored assessment of individual children rather than on gross
classifications of abuse experiences may aid the design of child
welfare services that better protect children’s mental health. Men-
tal health providers, in turn, ought to be cognizant of these patterns
and work with child welfare agencies to ensure that sexually
abused children are provided psychological care that is appropriate
to their level of need and that is sensitive to other significant
experiences in children’s lives that may contribute to their psy-
chopathology symptoms or risk.
References
Achenbach, T. M. (1993). Empirically based taxonomy: How to use
syndromes and profile types derived from the CBCL/4–18, TRF, and
YSR. Burlington: University of Vermont, Department of Psychiatry.
Avery, L., Massat, C. R., & Lundy, M. (2000). Posttraumatic stress and
mental health functioning of sexually abused children. Child and Ado-
lescent Social Work Journal, 17, 19–34.
Bagley, C., & Mallick, K. (2000). Prediction of sexual, emotional, and
physical maltreatment and mental health outcomes in a longitudinal
cohort of 290 adolescent women. Child Maltreatment, 5, 218–226.
Barnett, D., Manly, J. T., & Cicchetti, D. (1993). Defining child maltreat-
ment: The interface between policy and child research. In D. Cicchetti &
S. L. Toth (Eds.), Child abuse, child development, and social policy:
Advances in applied developmental psychology (Vol 8, pp. 7–73). Nor-
wood, NJ: Ablex.
Beitchman, J. H., Zucker, K. J., Hood, J. E., daCosta, G. A., & Akman, D.
(1991). A review of the short-term effects of child sexual abuse. Child
Abuse & Neglect, 15, 537–556.
Beichtman, J., Zucker, K., Hood, J., DaCosta, G., Akman, S., & Cassavia,
E. A. (1993). A review of the long-term effects of child sexual abuse.
Child Abuse & Neglect, 191, 101–117.
Berrick, J. D., Needell, B., Barth, R. P., & Jonson-Reid, M. (1998). The
tender years: Toward developmentally sensitive child welfare services
for very young children. New York: Oxford University Press.
Bolger, K. E., & Patterson, C. J. (2001). Pathways from child maltreatment
to internalizing problems: Perceptions of control as mediators and mod-
erators. Development and Psychopathology, 13, 913–940.
Briere, J. (1996). Trauma Symptom Checklist for children: Professional
manual. Odessa, FL: Psychological Assessment Resources.
Burns, B. J., Phillips, S. D., Wagner, H. R., Barth, R. P., Kolko, D. J.,
Campbell, Y., & Landsverk, J. (2004). Mental health need and access to
mental health services by youths involved with child welfare: A national
survey. Journal of the American Academy of Child and Adolescent
Psychiatry, 43, 960–970.
Cohen, J. A., & Mannarino, A. P. (1996). Factors that mediate treatment
outcome of sexually abused preschool children. Journal of the American
Academy of Child and Adolescent Psychiatry, 37, 44 –51.
Danielson, C. K., deArellano, M. A., Kilpatrick, D. G., Saunders, B. E., &
Resnick, H. S. (2005). Child maltreatment in depressed adolescents:
Differences in symptomatology based on history of abuse. Child Mal-
treatment, 10, 37–48.
Dong, M., Anda, R. F., Dube, S. R., Giles, W. H., & Felitti, V. J. (2003).
The relationship of exposure to childhood sexual abuse to other forms of
abuse, neglect, and household dysfunction during childhood. Child
Abuse & Neglect, 27, 625–239.
Drake, B. (1995). Associations between reporter type and assessment
outcomes in child protective services referrals. Children and Youth
Services Review, 17, 503–522.
Finkelhor, D., & Berliner, L. (1995). Research on the treatment of sexually
abused children: A review and recommendations. Journal of the Amer-
ican Academy of Child and Adolescent Psychiatry, 34, 1408–1423.
Fergusson, D. M. & Mullen, P. E. (1999). Childhood Sexual Abuse: An
Evidence Based Perspective (Developmental clinical psychology and
psychiatry series, Vol. 40). Thousand Oaks, CA: Sage.
Garland, A. F., Landsverk, J. L., Hough, R. L., & Ellis-MacLeod, E.
(1996). Type of maltreatment as a predictor of mental health service use
for children in foster care. Child Abuse & Neglect, 20, 675– 688.
Hurlburt, M. S., Leslie, L. K., Landsverk, J., Barth, R. P., Burns, B. J.,
Gibbons, R. D., et al. 2004 Contextual predictors of mental health
service use among a cohort of children open to child welfare. Archives
of General Psychiatry, 61, 1217–1224.
Jonson-Reid, M., Drake, B., Chung, S., & Way, I. (2003). Cross-type
recidivism among child maltreatment victims and perpetrators. Child
Abuse & Neglect, 27, 899 –917.
Kellogg, N. D., & Menard, S. W. (2003). Violence among family members
of children and adolescents evaluated for sexual abuse. Child Abuse &
Neglect, 27, 1367–1376.
Kendall-Tackett, K. A., Williams, L. M., & Finkelhor, D. (1993). Impact
of sexual abuse on children: A review and synthesis of recent empirical
studies. Psychological Bulletin, 113, 164 –180.
Kessler, R. C., Andrews, G., Mroczek, D., Ustun, T. B., & Wittchen, H. U.
(1998). The World Health Organization Composite International Diag-
nostic Interview—Short Form (CIDI–SF). International Journal of
Methods in Psychiatric Research, 7, 171–185.
Kilpatrick, D. G., Ruggiero, K. J., Acierno, R., Saunders, B. E., Resnick,
H. S., & Best, C. L. (2003). Violence and risk of PTSD, major depres-
sion, substance abuse/dependence, and comorbidity: Results from the
National Survey of Adolescents. Journal of Consulting and Clinical
Psychology, 71, 692–700.
Kolko, D. J., Barth, R. P., Burns, B. J., Zhang, J., Leslie, L., & Fairbank,
J. A. 2005. Posttraumatic stress in children receiving child welfare
services: A national sample of in-home and out-of-home care. Manu-
script submitted for publication.
Kovacs, M. (1992). Children’s depression inventory. North Tonawanda,
NY: Multi-Health Systems.
Lau, A. S., Leeb, R. T., English, D., Christopher-Gram, J., Briggs, E. C.,
480
MCCRAE, CHAPMAN, AND CHRIST
Page 13
Brody, K. E., & Marshall, J. M. (2005). What’s in a name? A compar-
ison of methods for classifying predominant type of maltreatment. Child
Abuse & Neglect, 29, 533–551.
Manly, J. T., Cicchetti, D., & Barnett, D. (1994). The impact of subtype,
severity, chronicity, and frequency of child maltreatment on social
competence and behavior problems. Development and Psychopathology,
6, 121–143.
McCrae, J. S., Barth, R. P., & the NSCAW Research Group. (2004).
Maltreatment types and risk associated with post-investigative child
welfare services. Manuscript submitted for publication.
McLeer, S. V., Dixon, J. F., Henry, D., Ruggiero, K., Escovitz, K., Niedda,
T., et al. (1998). Psychopathology in non-clinically referred sexually
abused children. Journal of the American Academy of Child and Ado-
lescent Psychiatry, 37, 1326 –1333.
Molnar, B. E., Buka, S. L., & Kessler, R. C. (2001). Child sexual abuse and
subsequent psychopathology: Results from the National Comorbidity
Survey. American Journal of Public Health, 91, 753–760.
Muthe´n, B., & Muthe´n, L. K. (2000). Integrating person-centered and
variable-centered analyses: Growth mixture modeling with latent trajec-
tory classes. Alcoholism Clinical and Experimental Research, 24, 882–
891.
NSCAW Research Group. (2002). Methodological lessons from the Na-
tional Survey of Child and Adolescent Well-Being: The first three years
of the USA’s first national probability study of children and families
investigated for abuse and neglect. Children and Youth Services Review,
24, 513–541.
Paradise, J. E., Rose, L., Sleeper, L. A., & Nathanson, M. (1994). Behavior,
family function, school performance, and predictors of persistent distur-
bance in sexually abused children. Pediatrics, 93, 452– 459.
Putnam, F. W. (1996). Special methods for trauma research with children.
In E. B. Carlson (Ed.), Trauma Research Methodology (pp. 153–173).
Lutherville, MD: Sidran.
Research Triangle Institute (2001). SUDAAN user’s manual (Version 8.0)
[Computer software]. Research Triangle Park, NC: Author.
Romano, E., & De Luca, R. V. (2001). Male sexual abuse: A review of
effects, abuse characteristics, and links with later psychological func-
tioning. Aggression and Violent Behavior, 6, 55–78.
Ruggiero, K. J., McLeer, S. V., & Dixon, J. F. (2000). Sexual abuse
characteristics associated with survivor psychopathology. Child Abuse
& Neglect, 24, 951–964.
Runyan, D. K. (1998). Prevalence, risk, sensitivity, and specificity: A
commentary on the epidemiology of child sexual abuse and the devel-
opment of a research agenda. Child Abuse & Neglect, 22, 493– 498.
Saunders, B. E., Berliner, L., & Hanson, R. F. (2001). Guidelines for the
psychosocial treatment of intrafamilial child physical and sexual abuse.
Final Draft Report. Charleston, SC: Author.
Straus, M. (1979). Measuring intrafamily conflict and violence: The Con-
flict Tactics (CT) scales. Journal of Marriage and the Family, 41,
75– 88.
Thornberry, T. P., Ireland, T. O., & Smith, C. A. (2001). The importance
of timing: The varying impact of childhood and adolescent maltreatment
on multiple problem outcomes. Development and Psychopathology, 13,
957–979.
Trickett, P. K., Noll, J. G., Reiffman, A., & Putnam, F. W. (2001). Variants
of intrafamilial sexual abuse experience: Implications for short-and
long-term development. Development and Psychopathology, 13, 1001–
1019.
Turner, H. A., Finkelhor, D., & Omrod, R. (2006). The effect of lifetime
victimization on the mental health of children and adolescents. Social
Science & Medicine, 62, 13–27.
U.S. Department of Health and Human Services, Administration on Chil-
dren, Youth, and Families (2002). Summary of the results of the 2001
and 2002 Child and Family Services Reviews. Retrieved from http://
www.acf.hhs.gov
U.S. Department of Health and Human Services, Administration for Chil-
dren and Families (2004). National Survey of Child and Adolescent
Well-Being: Children involved with the child welfare system (Wave 1).
Washington, DC: Author.
Waldfogel, J. (1998). The future of child protection. Cambridge, MA:
Harvard University Press.
Wolfe, D. A. (1999). Child abuse: Implications for child development and
psychopsychology (2nd ed.). Thousand Oaks, CA: Sage.
Received March 18, 2005
Revision received August 30, 2005
Accepted May 5, 2006
481
CSA, PSYCHOPATHOLOGY, AND SERVICES
Page 14
  • Source
    • "The number of adolescents who reported different types of maltreatment in the range of low to moderate–severe levels of severity was highest for Class 1, followed by Class 2, 3 and 4, respectively. Collectively, these findings highlight the role of multiple maltreatment and severity within each type of maltreatment (Finkelhor et al., 2009; Higgins & McCabe, 2001; Lau et al., 2005) leading to differences in distribution patterns or discrete latent-classes as found in the present study (McCrae, Chapman, & Christ, 2006; Pears et al., 2008). Noteworthy, is that more than 40% of the adolescents in Class 1 reported moderate–severe levels of abuse/neglect, 42.5% reported moderate–severe levels of sexual abuse and nearly a quarter reported low levels of abuse/neglect in Class 2, and more than 70% of the adolescents in Class 3 reported moderate–severe levels of neglect (Table 2). "
    [Show abstract] [Hide abstract] ABSTRACT: Objectives: The aims of the present study were, to identify discrete classes of adolescents based on their reporting of emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect of several levels of severity using a person-centered analytic approach (i.e., latent class analysis), and to compare the latent classes on 17 dimensions of personality pathology. It was hypothesized that based on types of maltreatment and severity levels within each type there would be discrete latent classes, and that classes of adolescents exposed to a larger number of maltreatment types with higher severity (i.e., moderate-severe) would report higher levels of personality pathology than adolescents in classes exposed to less types with less severity, after controlling for age and gender. Methods: Participants were 702 adolescents from Jammu, India (13-17 years, 41.5% females). The latent classes were based on three levels of severity for each type of maltreatment assessed via the Childhood Trauma Questionnaire (Bernstein et al., 2003). Results: Four distinct classes, namely, Moderate-severe abuse and physical neglect (Class 1), Low to moderate-severe abuse (Class 2), Moderate-severe neglect (Class 3), and Minimal abuse or neglect (Class 4) were found. Classes with higher percentages of adolescents reporting abuse and neglect with higher severity (Classes 1 and 2) reported higher levels of personality pathology than the other classes. Conclusions: There are distinct classes of adolescents’ identifiable based on levels of severity and types of abuse and neglect, which are differentially associated with specific dimensions of personality pathology. Implications for research and practice are discussed.
    Full-text · Article · Jun 2015 · Child Abuse & Neglect
  • Source
    • "Fit indices (seeTable 2) suggested that Akaike information criterion and adjusted Bayesian information criterion significantly improved as the number of profiles increased from the one-profile to the three-profile model, and they showed small improvements to the four-and five-profile models. In addition, the fourand five-profile models had one profile comprising less than 5% of the sample, suggesting the solution may be unstable (McCrae, Chapman, & Christ, 2006). Thus, the three-profile model was selected as the best fitting solution. "
    [Show abstract] [Hide abstract] ABSTRACT: This prospective longitudinal study examines the long-term influence of intimate partner violence (IPV) exposure in utero. We hypothesized that (a) prenatal IPV increases risk for internalizing and externalizing problems as well as for a profile of dysregulated cortisol reactivity, and (b) patterns of cortisol hyper- and hyporeactivity are differentially associated with internalizing and externalizing problems. The participants were 119 10-year-old children. Their mothers reported their IPV experiences and distress during pregnancy. Child and maternal reports of internalizing and externalizing problems as well as lifetime IPV exposure were obtained. Salivary cortisol was assessed at baseline, 20 min, and 40 min after challenge. The results partially supported our hypotheses: Exposure to IPV during pregnancy predicted child-reported internalizing and externalizing problems, mother ratings of child externalizing problems, and a profile of high cortisol secretion before and after stress challenge. The results were significant above and beyond the influence of maternal distress during pregnancy and IPV that occurred during the child's life. In addition, a profile of high cortisol secretion was associated with maternal reports of child internalizing behaviors. Findings support the growing consensus that prenatal stress can lead to lasting disruptions in adaptation and highlight the need for more longitudinal examinations of prenatal IPV exposure.
    Full-text · Article · Apr 2015 · Development and Psychopathology
  • Source
    • "and sexual abuse , and 93 . 48% experiencing both psy - chological abuse and neglect . These findings are similar to those in the Trickett et al . ( 2009 ) study , in which emotional abuse accompa - nied both physical abuse and neglect . Child abuse often occurs in families experiencing domestic violence or marital discord ( Burke et al . , 2011 ; McCrae et al . , 2006 ; Shipman , Rossman , & West , 1999 ) . The findings in this study are consistent with this research ; however , physical abuse was only related to caregiver domestic violence while sexual abuse was only associated with caregiver mental health problems , with over half of the children reporting both adversities , and neglect was associa"
    [Show abstract] [Hide abstract] ABSTRACT: Exposure to adverse childhood experiences (ACEs) such as child abuse and neglect impact a child's socioemotional development. Drawing from the methods employed in the Adverse Childhood Experiences (ACE; Felitti et al., ) Study, the present study utilized data from the National Survey of Child and Adolescent Well-Being to examine the prevalence of ACEs among children birth to 6 years, and the relationship of ACEs to emotional and behavioral outcomes 59 to 97 months after the close of investigation or assessment. Logistic regression also was used to examine the cumulative impact of ACEs on child behavior outcomes. By the age of 6, approximately 70% of children experienced three or more ACEs, and there were strong relationships between ACEs. Numerous ACEs were associated with long-term behavioral problems, and results supported a dose-response effect. Three or greater ACEs more than quadrupled the risk of experiencing internalizing problems, and almost quadrupled the risk of experiencing either externalizing or total problems at 59 to 97 months' postinvestigation. Based on these findings, it is crucial for both early screening/assessment and increased collaboration between child welfare and early intervention programs. © 2014 Michigan Association for Infant Mental Health.
    Full-text · Article · Nov 2014 · Infant Mental Health Journal
Show more