1. No category

This article was originally published in a journal published by
Elsevier, and the attached copy is provided by Elsevier for the
author’s benefit and for the benefit of the author’s institution, for
non-commercial research and educational use including without
limitation use in instruction at your institution, sending it to specific
colleagues that you know, and providing a copy to your institution’s
administrator.
All other uses, reproduction and distribution, including without
limitation commercial reprints, selling or licensing copies or access,
or posting on open internet sites, your personal or institution’s
website or repository, are prohibited. For exceptions, permission
may be sought for such use through Elsevier’s permissions site at:
http://www.elsevier.com/locate/permissionusematerial
Children and Youth Services Review
29 (2007) 941 – 960
py
www.elsevier.com/locate/childyouth
co
The unique impact of out-of-home placement and the
mediating effects of child maltreatment and
homelessness on early school success
John Fantuzzo ⁎, Staci Perlman
al
University of Pennsylvania, USA
on
Received 11 September 2006; received in revised form 3 November 2006; accepted 15 November 2006
Available online 8 February 2007
rs
Abstract
th
o
r's
pe
Increased national attention has underscored the importance of promoting educational well-being for
children who have been placed in out-of-home care. The present study, informed by a developmental
epidemiology framework, examined the unique impact of out-of-home placement and the mediating effects of
child maltreatment and homelessness on the academic achievement and school adjustment of an entire cohort
of second grade children in a large urban school district. Data on birth risks, placement history, child
maltreatment, and homelessness from birth through second grade were integrated across municipal agencies
for over 11,000 second grade students. Multiple Logistic Regression analyses demonstrate that children with a
history of out-of-home placement were at increased risk for poor literacy and science achievement controlling
for demographics and birth risks. These children also evidenced significantly higher levels of behavior
problems and school suspensions than children with no out-of-home placement history. Maltreatment and
homelessness were found to have significant mediating effects on the relationship between out-of-home
placement and children's educational well-being. Implications for policy and practice were discussed.
© 2007 Elsevier Ltd. All rights reserved.
Au
Keywords: Multiple Logistic Regression Analyses; Out-of-home placement; Maltreatment; Mediating effects
1. Introduction
No Child Left Behind legislation has pressed American public schools to ensure that all children
are meeting minimum academic standards by third grade (U.S. Department of Education, 2004).
⁎ Corresponding author. University of Pennsylvania, Graduate School of Education, 3700 Walnut Street, Philadelphia,
PA 19104-6216, USA. Tel.: +1 215 898 4790; fax: +1 215 573 2115.
E-mail address: [email protected] (J. Fantuzzo).
0190-7409/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.childyouth.2006.11.003
942
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
Au
th
o
r's
pe
rs
on
al
co
py
By setting the target at grade three, this legislation affirms the significance of early childhood and
the necessity of early identification and intervention for vulnerable young children. Recent
research has highlighted two major realities drawn from the early childhood research related to
educational well-being: one, multiple early risk factors have an adverse affect on later school
functioning, and two, the capacity of early intervention to increase favorable outcomes for
vulnerable children living in high risk environments (National Research Council, 1998, 2000,
2001).
Neurons to Neighborhoods, a major report by the National Research Council (2000)
documented the importance of early childhood experiences and the connection between
children's ability to successfully negotiate early developmental challenges and their future
development. Young children exposed to multiple early childhood risk factors are at increased
risk for not acquiring these developmental competencies. The report identified two types of risk
factors: biological and social. Biological risk factors include, for example, low birth weight and
premature birth. The second type, social risk, includes factors such as out-of-home placement,
homelessness, and child maltreatment. Each of these risk factors has been demonstrated to place
children at risk for poor academic outcomes (Huffman, Mehlinger, & Kerivan, 2000; McWayne,
Fantuzzo, & McDermott, 2004; Weiss & Fantuzzo, 2001). This work affirms the need for a
broader developmental approach to understanding risk.
The developmental epidemiology model provides a comprehensive conceptual framework for
examining the impact of biological and social risk factors on children's development (Costello &
Angold, 2000). The epidemiological aspects of this model emphasize the utility of a public health
model to collect population-based data. Population-based data can be used to understand both the
prevalence of various risk factors and how these risk factors relate to children's well-being. In
accordance with a public health model, these data should be collected by key frontline sentinels for
an entire population of children. Developmental epidemiology also draws from developmental
science in its recognition of the importance of a stable, nurturing environment for children's
development and its understanding that there are multiple pathways of development. Research has
demonstrated that children in stable, nurturing relationships are more likely to have their basic needs
met (such as shelter and food) and thus more likely to have enhanced cognitive, physical, and
emotional well-being (Berrick, Needell, Barth, & Jonson-Reid, 1998; Brazelton & Greenspan, 2000;
Goldstein, Solnit, Goldstein, & Freud, 1996). Additionally, research has shown that multiple risk and
protective factors influence children's developmental trajectory (Masten & Wright, 1998). This
brings to light the importance of identifying the multiple risk factors that many of these children are
exposed to, and understanding how they relate to children's developmental trajectories and overall
educational well-being. The developmental epidemiology model therefore, can be employed to
study significant risk factors that threaten children's opportunity to develop within a safe and stable
environment and to explore how these risks impact children's educational well-being.
The national child welfare system serves as a public surveillance system designed to identify
major social risks impacting children. Based on the understanding that children best thrive in the
context of a safe, stable, and nurturing family environment (Goldstein, Solnit, Goldstein, &
Freud, 1996), state and local child welfare systems are charged through federal legislation to
address risks that disrupt a child's family environment. Being placed outside the home in
substitute care arrangements, such as foster care, kinship care, etc. represents one type of risk that
the child welfare system monitors.
According to the most recent national data, approximately 290,000 children entered out-ofhome placement during Fiscal Year 1999 (United States Department of Health and Human
Services [USDHHS], 2003). This population is disproportionately comprised of children 5 years
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
943
Au
th
o
r's
pe
rs
on
al
co
py
of age or younger (USDHHS, 2003). A major policy report by the National Center for Children in
Poverty (2001) identified these young children as being “among the most vulnerable children in
the country.” Furthermore, a small body of research has documented the adverse impact of out-ofhome placement on young children's development and school-age children's academic
achievement and engagement.
Although research has identified the importance of developmentally appropriate child welfare
services for young children (Berrick, Needell, Barth, & Johnson-Reid, 1998), few empirical
studies have evaluated the impact of out-of-home placement on the overall well-being of children
ages 0–5 and their adjustment in elementary school. Four studies based on developmental
assessments conducted as part of developmental screening programs found that infants and
toddlers in out-of-home placement evidenced a high rate of developmental delays (Horwitz,
Simms, & Farrington, 1994; Klee, Kronstadt, & Zlotnick, 1997; Leslie, Gordon, Ganger, & Gist,
2002; Reams, 1999). An additional study found that infants and toddlers in out-of-home
placement disproportionately experienced language, cognitive or gross motor development
delays (Silver et al., 1999). Collectively, these studies document that the early development of
young children with out-of-home placement histories may be compromised even before they
enter school.
Concurrent and retrospective studies demonstrate that children with out-of-home placement
experiences are also at high risk of poor academic achievement while in school, and greater risk
for school drop out and unemployment. Children with out-of-home placement histories evidenced
severe delays in reading and were more likely to perform below grade level in math, language and
overall performance compared to their peers (Heath, Colton, & Aldgate, 1994; Mitic & Rimer,
2002; Stein, 1997; Zima et al., 2000). Furthermore, one study found that children in out-of-home
placement due to child maltreatment evidenced a long-term decline in school achievement (Heath
et al., 1994). Other studies document that children with a history of out-of-home placement were
at increased risk for suspensions and expulsions, grade retention, and drop-out (Dumaret, 1985;
Whiting-Blome, 1997; Zima, et al., 2000).
In addition to facing an increased risk of poor academic achievement, children with histories of
out-of-home placement are also at increased risk of poor academic adjustment (Colton & Heath,
1994; Evans, 2004; Fanshel & Shinn, 1978; Stein, 1997; Zima et al., 2000). A study conducted by
Canning (1974) found that children in out-of-home placement were more likely to evidence
“withdrawn”, “aggressive”, or “conformist” behaviors in the classroom. In a study of teachers'
assessments of children with a history of out-of-home placement, teachers rated children in foster
care as being less likely to engage in prosocial behaviors and more likely to have difficulty with
peer relationships (Stein, 1997). Similarly, Zima et al. (2000) found that 34% of children with outof-home placement experience had at least one classroom related behavioral problem. In a
longitudinal study conducted by Fanshel and Shinn (1978), children in out-of-home care were
found to have high rates of school absences. In a population-based study by Weiss and Fantuzzo
(2001), children in out-of-home placement were at 28% greater risk of having poor attendance
than children without a history of out-of-home placement. In sum, the research demonstrates that
out-of-home placement severely compromises children's educational well-being across academic
and behavioral adjustment outcomes.
This initial body of work underscores the importance of understanding the educational wellbeing of children with a history of out-of-home placement, however a critique of this literature
using the developmental epidemiology framework highlights several methodological and
conceptual limitations. This literature is particularly limited by small sample sizes, samples of
convenience, inadequate assessment of children's educational well-being, and failure to consider
944
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
Au
th
o
r's
pe
rs
on
al
co
py
the co-occurrence of other risk factors. Most of the literature addressing the relationship between
out-of-home placement and the educational well-being of children has been based on small
samples. Small sample sizes can limit empirical research by making them unsuitable for more
complex multivariate statistics. Multivariate statistics, such as linear and logistic regression,
permit the examination of the unique impact of each social risk factor while simultaneously
accounting for other known risk factors. This means it is possible to calculate the relative risk
associated with each risk factor, controlling for the influence of other risk factors the child may
have been exposed to. Of the out-of-home placement studies addressing academic achievement
and adjustment, only six had sample sizes large enough to employ complex multivariate models
(Evans, 2004; Fanshel & Shinn, 1978; Sawyer & Dubowitz, 1994; Weiss & Fantuzzo, 2001;
Whiting-Blome, 1997; Zima et al., 2000).
In addition to using small samples, many of the studies reviewed also utilized convenience
samples. Several studies of out-of-home placement were conducted as part of development
assessment or screening programs targeted at early identification of developmental delays
(Horwitz et al., 1994; Klee et al., 1997; Leslie et al., 2002; Silver et al., 1999). Findings from these
studies can only be generalized to the population of children who were permitted to participate in
the assessment programs or those referred to the researchers.
Some of the out-of-home placement studies utilized a multidimensional approach to assessing
children's educational well-being (i.e. they accounted for both academic achievement and
academic engagement), however many of these studies relied heavily on psychiatric checklists as
the means for evaluating children's academic engagement. Most of the studies employed
measures, such as the Child Behavior Checklist, to measure this competency (Fanshel & Shinn,
1978; McCue-Horwitz, Balestracci, & Simms, 2001; Stein, 1997; Zima et al., 2000). Prior
research has demonstrated that checklists such as these may not be appropriate for use with
diverse, low-income populations because they consider behaviors as narrow, one-dimensional,
parent-rating scales (Fantuzzo & Mohr, 2000).
Using the developmental epidemiology framework, a major critique of the out-of-home
placement literature is the failure of many studies to account for the co-occurrence of other social
risk factors. Although the out-of-home placement literature demonstrates an adverse relationship
between out-of-home placement and children's educational well-being, there has not been an
examination of how the impact of this risk factor might be mediated by the co-occurrence of other
risk factors. This limitation is notable because several studies have found that reason for entry into
out-of-home placement explains some of the variance in children's educational well-being
(Aldgate, Colton, Ghate, & Heath, 1992; Colton, Heath, & Aldgate, 1995; Heath et al., 1994) —
suggesting that children's experiences prior to entry into placement may mediate the relationship
between placement and educational well-being (Rutter, 2001; Wolkind & Rutter, 1973). This
finding becomes particularly salient in light of the known co-occurrence of out-of-home
placement, child maltreatment, and homelessness.
Several studies have documented the co-occurrence of out-of-home placement, homelessness,
and child maltreatment. One study found that 33% of homeless women had experienced some
form of out-of-home placement in their childhood (Zlotnick, Robertson, & Wright, 1999).
Another study by Park, Metraux, Brodbar, and Culhane (2004) found that 16% of homeless
children were in out-of-home placement either before or after their first shelter admission.
Additionally, several studies found that 12–15% of children experiencing child maltreatment
were placed in out-of-home care (Jonson-Reid, 2003; USDHHS, 2003). In light of the high cooccurrence of out-of-home placement with each of these risk factors, it is critical to explore how
this co-occurrence impacts children's educational well-being both uniquely and cumulatively.
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
945
co
py
The present study examines the unique risk associated with out-of-home placement across
multiple academic outcomes and the extent to which this risk is mediated by the presence of two
other social risk factors — child maltreatment and homelessness. It addresses the sampling
limitations of prior studies by utilizing population-level data collected by frontline sentinels for an
entire cohort of second grade children. Guided by a developmental epidemiology framework, the
study addressed four major research questions. First, what is the social risk history within a cohort
of second grade students in a large urban school district? Examination of social risk history
included out-of-home placement, child maltreatment and homelessness experiences. Second, what
is the unique impact of out-of-home placement on children's early academic achievement and
school adjustment at the end of second grade? A multivariate model was used to assess the effect of
out-of-home placement, controlling for demographics and birth risks on multiple academic and
behavioral outcomes. Third, what is the co-occurrence in the population of out-of-home placement
with child maltreatment and homelessness? Fourth, to what extent do child maltreatment and
homelessness function as mediating factors on the influence of out-of-home placement?
al
2. Method
on
2.1. Participants
pe
rs
This study was conducted in a large northeastern city with an entire cohort of second grade
children enrolled in the city's public school system. Participants included 11,835 students from the
population of 16,271 children who were: enrolled in second grade during the 2002–2003 academic
year, and born to mothers living in the city. Children were equally distributed between males (52%)
and females (48%), with an average age of 8.5 years (SD 0.52) at the end of second grade. Sixtyseven percent were African American, 15% Caucasian, 14% Hispanic, and 4% Asian.
2.2. Procedure
Au
th
o
r's
To develop an understanding of children's early childhood risks, a large integrated dataset was
created using citywide administrative data. Data for the study were obtained through the Kids
Integrated Data System (KIDS) (Fantuzzo, Culhane, & Hadley, 2005). KIDS is a database
infrastructure established by the University of Pennsylvania, under agreement with the
participating municipality. Its purpose is to support integrated database research to inform
practice and policies for children and youth. Participating agencies include the Department of
Public Health (Medicaid behavioral health, birth records, lead registry), the Department of
Human Services (child abuse and neglect, preventive services), the School District (attendance,
achievement, standardized testing, special education), Behavioral Health System (mental health
services) and the Office of Emergency Shelter and Services (public shelter use). A Memorandum
of Understanding between the City, State, and the University outlines the procedures under which
the data may be used. These procedures ensure confidentiality under the Health Insurance
Portability and Accountability Act (HIPAA) and Family Educational Rights and Privacy Act
(FERPA). KIDS employs advanced methods to ensure data quality and integrity. Complex
computer algorithms are used to match individuals and services across systems over time. Data
management includes reliability and validity auditing of data elements and the maintenance of
data standards.
For the present study, a linked dataset of the Department of Public Health (DPH), Department
of Human Services (DHS), Office of Emergency Shelter Services (OESS), and School District
946
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
on
al
co
py
was obtained. All identifying information, such as names, addresses, etc., was used solely for
matching purposes and the final data set was stripped of identifiers other than identification
numbers. After the matching was complete, individual level data on each child were then
extracted using identification numbers and appended to the core data set. Thus, a large data set
was formed containing birth records, homeless experience, child maltreatment and out-of-home
placement data, and school information for each child.
The matching process used to link the data into an integrated dataset was completed using a
personal computer and Microsoft SQL Server software (2005). Data from each of the
participating agencies was standardized prior to the matching process. Additionally, duplicate
records were eliminated from each dataset prior to matching. Matching algorithms were used to
link children across the datasets. Once the matching process was completed, observations for
which there was a possible false positive error were identified. These observations comprised less
than 1% of all matches in each dataset and were manually cross-referenced across each of the
datasets to ensure accuracy. The resulting integrated dataset consisted of 16,271 second grade
children. Of these children, 78.2% had complete birth records. To ensure that only children with
complete risk histories were included in the study, only the 78.2% (n = 16,271) of children with
complete birth records were included in analyses.
2.3. Measures
th
o
r's
pe
rs
2.3.1. Birth risks
Birth risks data were provided by the Department of Public Health. The variable ‘birth risks’
was defined as whether or not a child experienced at least one of the following risk factors:
inadequate prenatal care, premature birth, or low birth weight. Each of these risk factors was
recorded on the birth certificate and entered into a larger birth record database. Children identified
as having received no prenatal care, prenatal care only in the third trimester, or fewer than four
prenatal visits were considered to have received inadequate prenatal care. Those whose mothers
received more than four visits throughout the course of their pregnancies were identified as
having receiving adequate prenatal care. Similarly, children born at less than 36 weeks gestation
were considered to be premature. Those born after 36 weeks gestation were considered to be fullterm. Lastly, children experiencing low birth weight were identified through their birth record.
Typically, children weighing less than 2500 g are considered to have low birth weight. A binary
variable was created to determine the presence or absence of birth risks. Children experiencing
any one of these risks were considered to have experienced birth risks. Children experiencing
none were considered not at risk.
Au
2.3.2. Poverty
Children were defined as having had an experience of poverty if they received a free or
reduced school lunch. These data were recorded in a dataset maintained by the School District. A
binary variable was created to determine the presence or absence of poverty. Children identified
as qualifying for a free or reduced lunch either in second grade or in a prior year were considered
to have experienced poverty. Those who did not qualify or who were not identified within the
database were considered to be not at risk.
2.3.3. Out-of-home placement
Out-of-home placement was identified using data provided by DHS. DHS maintains a
database of all placement services that are paid for by the Agency. There are five types of out-of-
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
947
py
home placement services provided by DHS: kinship care, foster care, group home care,
institutional care, and supervised independent living. A binary variable was created to determine
the presence or absence of out-of-home placement experience. Children with a history of at least
one placement in kinship care, foster care, group home care, or non-homeless shelter institutional
care by the end of second grade were considered to have experienced the risk factor of out-ofhome placement. Those without a history of placement in at least one of those settings were
considered to be not at risk.
on
al
co
2.3.4. Child maltreatment
DHS also provided data on substantiated child maltreatment. DHS maintains a database
tracking system that archives each allegation of child maltreatment. Within the municipality,
substantiated child maltreatment is designated by a Child Protective Services (CPS) or General
Services Report (GPS) that is substantiated, indicated, or founded. A binary variable was created
to determine the presence or absence of child maltreatment. Children with a history of at least one
substantiated, founded, or indicated allegation of child maltreatment by the end of second grade
were considered to have experienced the risk factor of child maltreatment. Children without a
history of at least one substantiated, founded, or indicated allegation of child maltreatment were
considered not at risk.
pe
rs
2.3.5. Homeless experience
Homeless experience was defined as whether or not the child had ever been placed in a
homeless shelter. Information regarding children's homeless experiences was collected by both
OESS and DHS. A binary variable was created to determine the presence or absence of homeless
experience. If a child's parent was identified within the OESS database or a child was identified as
having been placed in a DHS-funded homeless shelter, the child was considered to have had a
homeless experience. If a child's parent was not identified within the OESS database and the child
was not identified as having been placed in a DHS-funded homeless shelter, then the child was
considered to not have had a homeless experience.
Au
th
o
r's
2.3.6. Academic achievement
All academic achievement data were provided by the School District. The Dynamic Indicators
of Basic Early Literacy Skills (DIBELS; Good & Kaminski, 2002) is a standardized measurement
system of early literacy development. It is designed to identify children with early reading
difficulty and to monitor progress within the curriculum. The DIBELS contain a set of 10 brief
tests individually administered by educators to children in Kindergarten through third grade. Each
subtest is a short (1–3 min) fluency measure based on the early literacy domains defined in both
the National Reading Panel (National Institute of Child Health and Human Development, 2000b)
and National Research Council (1998) reports. Subtests include measures of children's
phonological awareness, knowledge of alphabetic print, and language development. Authors
recommend a different combination of subtests, depending on children's developmental level
(i.e., grade), to be administered at different times throughout the academic year. The subtest used
to assess second grade children at the end of the school year is Oral Reading Fluency (ORF). This
brief test allows for a quick identification of children's accuracy and fluency while reading aloud.
Unpublished technical reports for the DIBELS provide initial reliability and validity estimates for
the subtest used in second grade (Good, Wallin, Simmons, Kame'enui, & Kaminski, 2002).
Criterion validity for the ORF subtest ranged from .52 to .91 (Good et al., 2002). For the present
study, DIBELS scores were dichotomized at one standard deviation below the mean. Children
948
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
Au
th
o
r's
pe
rs
on
al
co
py
scoring less than one standard deviation below the mean were coded as ‘at risk’ for inadequate
school performance. Children scoring at or above one standard deviation below the mean were
coded as having adequate school achievement.
The Developmental Reading Assessment (DRA; Pearson Learning Group, 2003) is an
individually administered test of reading accuracy, fluency, and comprehension. It is appropriate
for use with children in grades kindergarten through third grade, and can be administered and
interpreted by classroom teachers in approximately 10 to 20 min. In a one-to-one format, students
read from texts within each level, progressing until they are unable to meet accuracy and
comprehension thresholds. The overall score on the DRA is an Instructional Reading Level — or
the level at which the student can engage in teacher-instructed text. Across all grade levels and
reading levels of the DRA, test–retest reliability estimates range from .91 to .99, and inter-rater
reliability estimates range from .74 to .80. Criterion-related construct validity has also been
established, with coefficients ranging from .65 to .84 when compared to scores on other nationally
standardized measures of early reading ability. For the present study, DRA scores were
dichotomized at the 15th percentile. Children scoring below the fifteenth percentile were coded as
‘at risk’ for inadequate school performance. Children scoring at or above the 15th percentile were
coded as having adequate school achievement.
The TerraNova, Second Edition (CTB/McGraw–Hill, 1997) is a group-administered
achievement test considered to be among the most reliable and valid of all standardized
achievement tests; it is also known as the California Achievement Tests, Sixth Edition. Standard
scores are provided across three subtests related to reading: reading, vocabulary, and language.
Standard scores were also provided for math and science subtests. The TerraNova was nationally
standardized on a stratified sample of 114,312 students (grades 1–12) from 778 school districts
during the fall of 1999 and another 149,798 students (grades K-12) in the spring of 2000.
Stratification variables included geographic region, urbanicity, socioeconomic status, and special
needs. The TerraNova demonstrates acceptable internal consistency, with Kuder–Richardson
Formula 20 coefficients for all subtests and total scores ranging from the mid .80s to .90s.
Extensive validity work has been conducted on the TerraNova. Items were carefully reviewed to
ensure adequate content validity, comparisons with the Test of Cognitive Skills, Second Edition
and with InView (CTB/McGraw–Hill, 2001) indicate evidence of construct validity, and
correlations between subtests and total scores support criterion-related validity. Further, the
publishers plan to correlate the TerraNova with the National Assessment of Educational Progress
(NAEP), the Third International Mathematics and Science Study (TIMSS), and the SAT and ACT.
Based on similar studies relating the California Achievement Tests with the SAT and ACT, the
publishers expect strong relationships. For the present study, TerraNova scores on each of the
subtests were dichotomized at the 15th percentile. Children scoring below the fifteenth percentile
were coded as ‘at risk’ for inadequate school performance. Children performing at or above the
15th percentile were coded as having adequate school achievement.
2.3.7. School adjustment
All academic adjustment data were provided by the School District. The Work Habits
Performance Assessment is a teacher evaluation of children's learning behaviors within the
classroom. There are eleven items rated on a scale ranging from 1 (Improvement Needed) to 3
(Competent). These items included, for example, “follows directions”, “asks for help when
necessary”, and “contributes information to class discussions”. Performance scores across all
eleven variables were averaged to create composite scores. Each composite score was then
standardized using the entire cohort of second grade performance assessments (N = 16,271).
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
949
2.4. Data analysis
r's
pe
rs
on
al
co
py
Internal consistency was demonstrated for the Work Habits (r = .90, p b .001). Criterionreferenced validity for Work Habits was established with the Learning Behaviors Scale (LBS;
Stott, McDermott, Green, & Francis, 1988) (r = .48, p b .0001). For the present study, Work Habits
scores were dichotomized at the 15th percentile. Children scoring below the fifteenth percentile
were coded as ‘at risk’ for inadequate academic engagement. Children scoring at or above the
15th percentile were coded as having adequate academic adjustment.
The Social Skills Performance Assessment is similar in format to the Work Habits
Performance Assessment and consists of nine items. Items included, for example, ‘works
cooperatively with others’, ‘displays a positive attitude’ and ‘displays appropriate behavior in
work and play’. Performance scores across all nine variables were averaged to create composite
scores. Each composite score was then standardized using the entire cohort of second grade
performance assessments (N = 16,271). Internal consistency was demonstrated for the Social
Skills Performance Assessment (r = .95, p b .001). Convergent and divergent validity for the
Social Skills Performance Assessment was established with the Adjustment Scales for Children
and Adolescents (ASCA; McDermott et al., 1999) in relation to underactive and overactive social
problems in the classroom (r = .18–.70). For the present study, Social Skills scores were
dichotomized at the 15th percentile. Children scoring below the fifteenth percentile were coded as
‘at risk’ for inadequate academic engagement. Children scoring at or above the 15th percentile
were coded as having adequate academic adjustment.
Attendance data were obtained from the School District's computerized records. A binary
variable was constructed to indicate children with poor attendance. To create a dichotomous
variable, the percentage of days absent were calculated for each student and divided into quartiles.
Attendance was coded as poor if absentees fell into the highest quartile and low if it fell into the
lowest three quartiles.
Similar to attendance, grade retention data were obtained from the School District's
computerized records. Children were identified as experiencing grade retention if they were
enrolled in second grade for both the 2002 and 2003 academic school years were identified as
experiencing grade retention.
Au
th
o
2.4.1. Descriptive picture of out-of-home placement and other social risk factors
The first set of analyses was conducted to develop an understanding of the demographic
characteristics and prevalence of history of out-of-home placement for a cohort of second grade
children. Frequency analyses and descriptive statistics were used to determine the prevalence of
these experiences within the entire cohort and to provide a descriptive picture of children with a
history of out-of-home placement. The second set of analyses explored the co-occurrence of outof-home placement with child maltreatment and homelessness within a cohort of second grade
children. Frequency analyses were used to determine the prevalence of these experiences within
the population of children with a history of out-of-home placement.
2.4.2. Unique impact of out-of-home placement on educational well-being
Multiple logistic regression was used to examine the association of out-of-home placement on
multiple outcomes of second grade academic achievement and academic adjustment, while
controlling for age, race, birth risks and poverty. This multivariate statistical technique was
chosen because it is frequently used in epidemiology research to assess the unique impact of
targeted risks on individual outcomes (Scott, Mason, & Chapman, 1999). It produces odds ratio
950
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
py
that quantify the magnitude of risk associated with each risk variable for each outcome (Nash &
Bowen, 2002) while controlling for the influence of other variables (e.g. demographics and other
risk factors). The overall chi-square statistic was examined to determine if the model was
significant and therefore whether the individual Wald chi-squares could be examined. For each
significant Wald chi-square, the odds ratio was inspected to assess its relative importance for the
outcome variable. The odds ratio is more easily interpreted as the degree of risk exerted by the
independent variables on the dependent variable.
rs
on
al
co
2.4.3. Child maltreatment and homelessness as mediating variables
Mediational analyses were conducted to explore the extent to which child maltreatment and
homelessness mediated the relationship between out-of-home placement and educational
outcomes. These analyses were conducted in steps outlined by Baron and Kenny (1986). First,
from the analyses conducted above, it was determined if out-of-home placement was significantly
related to each of the dependent variables. If statistical significance was found, the second step
involved determining if there were significant relationships between out-of-home placement and
each of the hypothesized mediating variables (child maltreatment and homelessness). Third,
where relationships were significant, logistic regressions were conducted to examine the extent to
which child maltreatment and/or homelessness mediated the relationship between out-of-home
placement and each of the outcome variables. For this last step, if the statistical significance and
magnitude of the odds ratios for out-of-home placement diminished upon introduction of each of
the hypothesized mediating variables, a mediating relationship was determined to exist. Findings
for the third and fourth sets of analyses are presented in the context of mediational analyses.
pe
3. Results
3.1. Prevalence of out-of-home placement and other social risk factors
Au
th
o
r's
Frequency analyses and descriptive statistics were used to develop a descriptive picture of
children with a history of out-of-home placement within a cohort of second grade children.
Frequency analyses revealed that approximately 3% of children in the second grade cohort had a
history of out-of-home placement. Eighty-seven percent of children with a history of out-of-home
placement were African American, 4% Caucasian, 8% Hispanic, and 1% Asian. There was a
slightly higher percentage of boys (58%) than girls with a history of out-of-home placement.
Additionally, the average age of first placement in out-of-home care for this group of children was
approximately 3.5 years. Over 50% of the children experienced more than one placement and on
average, they remained in care for just over 2 years (2.1).
Descriptive statistics were also used to determine the co-occurrence of out-of-home placement
with child maltreatment and homelessness. Frequency analyses revealed a high co-occurrence of
out-of-home placement with child maltreatment and/or homelessness. Within the population of
children with a history of out-of-home placement, approximately 35% had a history of
substantiated child maltreatment by the end of second grade. Over 70% of the population had a
history of homelessness by the end of second grade.
3.1.1. Unique risk of out-of-home placement
The unique risk of the independent variable, history of out-of-home placement, on academic
achievement and school adjustment was evaluated. Score statistics indicated that child
demographics, birth risks, poverty, and out-of-home placement variables were significantly
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
951
r's
pe
rs
on
al
co
py
related to children's second grade academic achievement. Table 1 presents the odds ratios and
probability levels for the independent effects of each demographic characteristic, birth risks,
poverty, and out-of-home placement on each academic achievement outcome. Out-of-home
placement was significantly related to children's performance on three of the literary assessments
and the second grade science assessment. Similar patterns of risk emerged for the following three
literacy assessments: DIBELS, Score (7, N = 9457) = 195.43, p b .0001; Language, Score (7,
N = 11,191) = 340.25, p b .0001; and Reading, Score (7, N = 11,191) = 307.78, p b .0001. Boys,
children of African American and Hispanic backgrounds, children living in low-income
households, children with birth risks and children with a history of out-of-home placement were
all associated with increased risk for poor outcomes in early literacy. Children with a history of
out-of-home placement, children of African American, Hispanic, and Asian backgrounds,
children living in low-income households, and children with birth risks, were all at increased risk
of poor achievement on the second grade standardized Science assessment, Score (7, N = 11,145) =
186.20, p b .0001. The first criterion for mediation was not met for the following academic
achievement outcomes: Vocabulary, Score (7, N = 11,088) = 282.27, p b .0001; Math, Score (7,
N = 11,258)= 266.61, p b .0001; and DRA, Score (7, N = 10,089) = 230.37, p b .0001. For each of
these outcomes, boys, children of African American and Hispanic backgrounds, children living in
low-income homes and birth risks all evidenced increased risk for poor outcomes however out-ofhome placement did not significantly predict these outcomes.
The unique risk of out-of-home placement on school adjustment was also assessed. Table 2
presents the odds ratios and probability levels for the independent effects of each demographic
characteristic, birth risks, poverty and out-of-home placement on each of the school adjustment
outcomes. The first criterion for mediation analyses was met for second grade Work Habits,
Social Skills, and School Suspension. Second grade Work Habits, Score (7, N = 10,296) = 440.13,
p b .0001, and second grade Social Skills, Score (7, N = 10,296) = 592.79, p b .0001 evidenced
similar patterns of risks. Boys, children of African American and Hispanic backgrounds, children
living in low-income homes, birth risks, and out-of-home placement all significantly predicted to
increased risk for poor Work Habits and Social Skills. Additionally, boys, children of African
American backgrounds, children living in low-income homes, birth risks, and out-of-home
placement all predicted to increased risk for School Suspension, Score (7, N = 11,771) = 313.63,
th
o
Table 1
Odds ratios and probability levels for independent effects of out-of-home placement on academic performance
DIBELS
DRA
TN Language TN Reading TN Vocabulary TN Math TN Science
Child characteristics
Gender (male)
African American
Hispanic
Asian
Poverty
Birth risks
Out-of-home placement
2.02⁎⁎⁎⁎
1.37⁎
2.35⁎⁎⁎⁎
.52
1.79⁎⁎⁎⁎
1.44⁎⁎⁎⁎
1.84⁎⁎⁎⁎
1.81⁎⁎⁎⁎
1.59⁎⁎⁎
1.89⁎⁎⁎⁎
.79
1.48⁎⁎⁎⁎
1.34⁎⁎⁎⁎
1.34
1.84⁎⁎⁎⁎
2.32⁎⁎⁎⁎
2.31⁎⁎⁎⁎
.83
1.63⁎⁎⁎⁎
1.27⁎⁎⁎⁎
1.57⁎⁎⁎⁎
Au
Risk factor/covariate
1.68⁎⁎⁎⁎
2.09⁎⁎⁎⁎
2.16⁎⁎⁎⁎
.78
1.65⁎⁎⁎⁎
1.36⁎⁎⁎⁎
1.35⁎⁎⁎⁎
1.81⁎⁎⁎⁎
1.79⁎⁎⁎⁎
2.18⁎⁎⁎⁎
.73
1.49⁎⁎⁎⁎
1.35⁎⁎⁎⁎
1.30
1.26⁎⁎⁎⁎
3.45⁎⁎⁎⁎
3.03⁎⁎⁎⁎
.70
1.38⁎⁎⁎⁎
1.41⁎⁎⁎⁎
1.04
1.04
3.51⁎⁎⁎⁎
3.30⁎⁎⁎⁎
1.74⁎⁎
1.26⁎⁎⁎⁎
1.23⁎⁎⁎
1.35⁎
Note. Score statistics for the above models were as follows: DIBELS = (7, N = 9457) = 195.43, p b .0001; DRA =
(7, N = 10,089) = 230.37 p b .0001; TN Language = (7, N = 11,191) = 340.25, p b .0001; TN Reading = (7, N = 11,191) =
307.78, p b .0001; TN Vocabulary = (7, N = 11,088) = 282.27, p b .0001; TN Math = (7, N = 11,258) = 266.61, p b .0001; and
TN Science = (7, N = 11,145) = 186.20, p b .0001.
Significance is based on Wald chi-square statistic.
⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001, ⁎⁎⁎⁎p b .0001.
952
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
Table 2
Odds ratios and probability levels for independent effects of out-of-home placement on school adjustment
Work habits
Social skills
Attendance
Suspension
Retention
Child characteristics
Gender (male)
African American
Hispanic
Asian
Poverty
Birth risks
Out-of-home
2.63⁎⁎⁎⁎
1.82⁎⁎⁎⁎
1.55⁎⁎⁎⁎
.24⁎⁎⁎⁎
1.56⁎⁎⁎⁎
1.25⁎⁎⁎⁎
1.68⁎⁎⁎
2.64⁎⁎⁎⁎
3.32⁎⁎⁎⁎
1.86⁎⁎⁎⁎
.19⁎⁎⁎⁎
1.36⁎⁎⁎⁎
1.15⁎⁎
1.80⁎⁎⁎⁎
1.29⁎⁎⁎⁎
1.46⁎⁎⁎⁎
1.17
.17⁎⁎⁎⁎
1.95⁎⁎⁎⁎
1.20⁎⁎
1.26
3.77⁎⁎⁎⁎
2.81⁎⁎⁎⁎
1.45
.26⁎
1.46⁎⁎⁎⁎
1.08
1.93⁎⁎⁎
1.33⁎⁎⁎
2.19⁎⁎⁎⁎
2.24⁎⁎⁎⁎
.44
1.81⁎⁎⁎⁎
1.30⁎⁎
.99
co
py
Risk factor/covariate
al
Note. Score statistics for the above models were as follows: Work Habits= (7, N = 10,296) = 440.13, p b .0001; Social skills =
(7, N = 10,296)= 592.79, p b .0001; Attendance = (7, N = 10,590)= 219.81, p b .0001; Suspension= (7, N = 11,771) = 313.63,
p b .0001; and Retention= (7, N = 11,771)= 128.99, p b .0001.
Significance is based on Wald chi-square statistic.
⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001, ⁎⁎⁎⁎p b .0001.
3.2. Mediational analyses
pe
rs
on
p b .0001). Although boys, children of African American background, children living in lowincome homes, and birth risks, all predicted to an increase in risk for poor school attendance,
Score (7, N = 10,590) = 219.81, p b .0001, out-of-home placement was not a significant predictor
and therefore the first criterion for mediation was not met. Similarly, for grade retention, Score (7,
N = 11,771) = 128.99, p b .0001, although boys, children of African American and Hispanic
backgrounds, living in low-income homes, and birth risks all predicted to an increase in risk, since
out-of-home placement was not significant, the first criterion for mediation was not met.
th
o
r's
3.2.1. Unique impact of out-of-home placement on the mediator variables
Consistent with Baron and Kenny's (1986) model for testing for mediation effects, the unique
relationship between out-of-home placement and each of the hypothesized mediating variables,
child maltreatment and homelessness, was evaluated. Findings indicated that out-of-home
placement was significantly associated with child maltreatment, Score (7, N = 11,771) = 343.45,
p b .0001 and homelessness, Score (7, N = 11,771) = 1102.43, p b .0001). As such, the second
criterion for mediation was met.
Au
3.2.2. Mediational impact of child maltreatment
Child maltreatment and out-of-home placement were entered simultaneously into a logistic
model for children's performance on the DIBELS, Language, Reading and Science (Table 3).
Similar patterns of results were found for two of the literacy assessments, DIBELS, Score (8,
N = 9457) = 227.59, p b .0001 and Language, Score (8, N = 11,191) = 382.33, p b .0001. Boys,
children of African American and Hispanic backgrounds, child living in low-income homes, birth
risks, out-of-home placement and child maltreatment all placed children at increased risk of poor
performance on each of these two performance assessments. Although out-of-home placement
remained statistically significant, its relationship to each of the two outcomes was substantially
attenuated by the presence of child maltreatment. Similar patterns of risk emerged for the second
grade standardized reading assessment, Score (8, N = 11,191) = 332.56, p b .0001 and science,
Score (8, N = 11,145) = 213.58, p b .0001). For reading, boys, children of African American and
Hispanic backgrounds, children living in low-income homes, birth risks, and child maltreatment
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
953
Table 3
Odds ratios and probability levels for mediation effects of child maltreatment on the relationship between out-of-home
placement and academic performance
DIBELS
TN Language
TN Reading
TN Science
Child characteristics
Gender (male)
African American
Hispanic
Asian
Poverty
Birth risks
Out-of-home placement
Child maltreatment
2.01⁎⁎⁎⁎
1.35⁎
2.39⁎⁎⁎⁎
.54
1.71⁎⁎⁎⁎
1.41⁎⁎⁎⁎
1.61⁎
1.87⁎⁎⁎⁎
1.83⁎⁎⁎⁎
2.31⁎⁎⁎⁎
2.33⁎⁎⁎⁎
.86
1.58⁎⁎⁎⁎
1.25⁎⁎⁎⁎
1.39⁎
1.67⁎⁎⁎⁎
1.67⁎⁎⁎⁎
2.07⁎⁎⁎⁎
2.17⁎⁎⁎⁎
.80
1.60⁎⁎⁎⁎
1.34⁎⁎⁎⁎
1.23
1.49⁎⁎⁎⁎
1.04
3.49⁎⁎⁎⁎
3.32⁎⁎⁎⁎
1.78⁎⁎
1.22⁎⁎⁎
1.21⁎⁎⁎
1.22
1.52⁎⁎⁎⁎
co
py
Risk factor/covariate
on
al
Note. Score statistics for the above models were as follows: DIBELS = (8, N = 9457) = 227.59, p b .0001; TN Language =
(8, N = 11,191) = 382.33, p b .0001; TN Reading = (8, N = 11,191) = 332.56, p b .0001; and TN Science = (8, N = 11,145) =
213.58, p b .0001.
Significance is based on Wald chi-square statistic.
⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001, ⁎⁎⁎⁎p b .0001.
r's
pe
rs
placed children at greater risk for poor performance. For science, children of African American
and Hispanic backgrounds, children living in low-income homes, birth risks, and child
maltreatment placed children at increased risk for poor performance. Notably, when child
maltreatment was entered into the model, out-of-home placement no longer had a unique impact
on reading and science outcomes. These findings demonstrated a partial mediation of child
maltreatment on the relationship between out-of-home placement and children's performance on
the DIBELS and a standardized language assessment and a perfect mediation of child
maltreatment on the relationship between out-of-home placement and second grade standardized
assessments of reading and science.
The mediational effect of child maltreatment on the relationship between out-of-home
placement and children's school adjustment was also evaluated (Table 4). Similar patterns
emerged for Social Skills, Score (8, N = 10,296) = 654.03, p b .0001 and Work Habits, Score
th
o
Table 4
Odds ratios and probability levels for mediation effects of child maltreatment on the relationship between out-of-home
placement and school adjustment
Work habits
Social skills
Suspension
Child characteristics
Gender (male)
African American
Hispanic
Asian
Poverty
Birth risks
Out-of-home placement
Child maltreatment
2.63⁎⁎⁎⁎
1.81⁎⁎⁎⁎
1.58⁎⁎⁎⁎
.25⁎⁎⁎⁎
1.48⁎⁎⁎⁎
1.22⁎⁎⁎
1.43⁎
2.02⁎⁎⁎⁎
2.64⁎⁎⁎⁎
3.31⁎⁎⁎⁎
1.89⁎⁎⁎⁎
.20⁎⁎⁎⁎
1.30⁎⁎⁎⁎
1.13⁎
1.56⁎⁎⁎
1.88⁎⁎⁎⁎
3.76⁎⁎⁎⁎
2.77⁎⁎⁎⁎
1.48⁎
.27⁎
1.35⁎⁎⁎
1.04
1.54⁎
2.44⁎⁎⁎⁎
Au
Risk factor/covariate
1.77⁎⁎
Note. Score statistics for the above models were as follows: Work habits = (8, N = 10,296) = 516.00, p b .0001; Social
skills = (8, N = 10,296) = 654.03, p b .0001; and Suspension = (8, N = 11,771) = 390.34, p b .0001.
Significance is based on Wald chi-square statistic.
⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001, ⁎⁎⁎⁎p b .0001.
954
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
co
py
(8, N = 10,296) = 516, p b .0001. In both cases, boys, children of African American and Hispanic
backgrounds, children living in low-income homes, birth risks, out-of-home placement and child
maltreatment were found to place children at increased for poor social skills and work habits. The
presence of child maltreatment in the model substantially minimized the impact of out-of-home
placement on each of the outcomes — demonstrating a partial mediation effect. The pattern of
risk for School Suspension, Score (8, N = 11,771) = 390.34, p b .0001 was similar except that birth
risks did not place children at increased risk for school suspensions. As with social skills and work
habits, child maltreatment was evidenced to be a partial mediator for the relationship between outof-home placement and school suspensions — with the effect of out-of-home placement
diminishing with the addition of child maltreatment.
r's
pe
rs
on
al
3.2.3. Mediational impact of homelessness
The mediational impact of homelessness on the relationship between out-of-home placement
and each of the academic achievement outcomes was also assessed (Table 5). Homelessness and
out-of-home placement were entered simultaneously into a logistic regression model for
children's performance on the DIBELS, Language, Reading and Science. A similar pattern of risk
emerged for children performance on the second grade standardized language, Score (8,
N = 11,191) = 389.99, p b .0001 and reading, Score (8, N = 11,191) = 338.20, p b .0001, assessments. Boys, children of African American and Hispanic backgrounds, children living in lowincome homes, birth risks, and homelessness all evidenced poor performance on each of these
academic performance measures. Children of African American, Hispanic, and Asian
backgrounds, children living in low-income homes, birth risks, and homelessness were at
increased risk for poor performance on a second grade science assessment, Score (8, N = 11,145) =
194.90, p b .0001. Notably, once homelessness was entered into the model out-of-home
placement no longer had a unique impact on each of the aforementioned outcomes. Last,
homelessness partially mediated the relationship between out-of-home placement and children's
performance on the DIBELS, Score (8, N = 9457) = 204.49, p b .0001. Boys, children of Hispanic
backgrounds, children living in low-income homes, birth risks, out-of-home placement and
homelessness, all predicted to poor performance on the DIBELS.
th
o
Table 5
Odds ratios and probability levels for mediation effects of homelessness on the relationship between out-of-home
placement and academic performance
TN DIBELS
TN Language
TN Reading
TN Science
Child characteristics
Gender (male)
African American
Hispanic
Asian
Poverty
Birth risks
Out-of-home placement
Homelessness
2.03⁎⁎⁎⁎
1.27
2.33⁎⁎⁎⁎
.53
1.73⁎⁎⁎⁎
1.41⁎⁎⁎⁎
1.58⁎
1.36⁎⁎
1.85⁎⁎⁎⁎
2.09⁎⁎⁎⁎
2.28⁎⁎⁎⁎
.84
1.54⁎⁎⁎⁎
1.24⁎⁎⁎
1.27
1.58⁎⁎⁎⁎
1.69⁎⁎⁎⁎
1.92⁎⁎⁎⁎
2.14⁎⁎⁎⁎
.79
1.58⁎⁎⁎⁎
1.33⁎⁎⁎⁎
1.14
1.43⁎⁎⁎⁎
1.04
3.37⁎⁎⁎⁎
3.29⁎⁎⁎⁎
1.75⁎⁎
1.23⁎⁎⁎
1.21⁎⁎⁎
1.24
1.21⁎
Au
Risk factor/covariate
Note. Score statistics for the above models were as follows: DIBELS =(8, N=9457) =204.49, pb .0001; TN Language =(8,
N =11,191) =340.25, pb .0001; TN Reading=(8, N =11,191) = 338.20, p b .0001; and TN Science=(8, N =11,145) =194.90,
p b .0001.
Significance is based on Wald chi-square statistic.
⁎p b .05, ⁎⁎p b .01, ⁎⁎⁎p b .001, ⁎⁎⁎⁎p b .0001.
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
955
Table 6
Odds ratios and probability levels for mediation effects of homelessness on the relationship between out-of-home
placement and school adjustment
Work habits
Social skills
Suspension
Child characteristics
Gender (Male)
African American
Hispanic
Asian
Poverty
Birth risks
Out-of-home placement
Homelessness
2.65⁎⁎⁎⁎
1.65⁎⁎⁎⁎
1.54⁎⁎⁎⁎
.25⁎⁎⁎⁎
1.49⁎⁎⁎⁎
1.22⁎⁎⁎
1.39⁎⁎⁎⁎
1.50⁎⁎⁎⁎
2.68⁎⁎⁎⁎
2.92⁎⁎⁎⁎
1.85⁎⁎⁎⁎
.20⁎⁎⁎⁎
1.27⁎⁎⁎⁎
1.11
1.40⁎
1.72⁎⁎⁎⁎
3.80⁎⁎⁎⁎
2.45⁎⁎⁎⁎
1.44
.26⁎⁎⁎⁎
1.35⁎⁎⁎⁎
1.03
1.49⁎
1.76⁎⁎⁎⁎
1.77⁎⁎
co
py
Risk factor/covariate
al
Note. Score statistics for the above models were as follows: Work habits = (8, N = 10,296) = 474.79, p b .0001; Social
skills = (8, N = 10,296) = 662.18, p b .0001; and Suspension = (8, N = 11,771) = 353.12, p b .0001.
Significance is based on Wald chi-square statistic.
⁎p b .05, ⁎⁎pb.01, ⁎⁎⁎pb.001, ⁎⁎⁎⁎pb.0001.
r's
pe
rs
on
The mediational effect of homelessness on the relationship between out-of-home placement
and children's school adjustment outcomes was also assessed (Table 6). Homelessness functioned
as partial mediator for the relationship between out-of-home placement and each of the school
adjustment outcomes. Boys, children of African American and Hispanic background, children
living in low-income homes, birth risks, out-of-home placement and homelessness were all at
increased risk for poor work habits, Score (8, N = 10,296) = 474.79, p b .0001. A similar pattern
emerged for social skills, Score (8, N = 10,296) = 662.18, p b .0001, except that the presence of
birth risks did not place children at increased risk of poor social skills. Last, boys, children of
African American backgrounds, children living in low-income homes, birth risks, out-of-home
placement and homelessness all placed children at increased risk of school suspension, Score (8,
N = 11,771) = 353.12, p b .0001. In sum, the unique risk associated with out-of-home placement
on children's school adjustment was significantly diminished once homelessness was entered into
the model.
th
o
4. Discussion
Au
Informed by a developmental epidemiology framework, the primary aim of this study was to
conduct a multivariate, population-based investigation of the prevalence and unique impact
of out-of-home placement on young urban children. This research involved three major foci:
(1) providing a descriptive picture of children with a history of out-of-home placement in an entire
cohort of public school second grade students in a large urban city; (2) determining the impact of
out-of-home placement on multiple academic and social adjustment outcomes controlling for
child demographics and birth risks; and (3) examining the hypothesized mediating effect of child
maltreatment and homelessness on the impact of out-of-home placement on school outcomes.
The study indicated that by the end of second grade nearly 3% of children in the cohort had a
history of out-of-home placement. On average children in this cohort with placement histories had
entered care by the time they were 4 years old, and approximately 75% of children with placement
histories had entered care by the time they were five. While prior studies have focused primarily
on the educational well-being of school age children, findings from this study are particularly
salient because they indicate that most of the children with placement histories had entered
956
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
Au
th
o
r's
pe
rs
on
al
co
py
placement prior to reaching school age. These findings illustrate the need for developing an
understanding of how early out-of-home placement experiences impact young children's school
readiness.
Findings demonstrate that children with a history of out-of-home placement were at increased
risk for poor performance on second grade standardized literacy and science assessments,
evidenced greater classroom behavior problems and experienced more school suspensions than
their peers. These findings corroborate with several studies that found that out-of-home placement
has an adverse relationship with academic outcomes and school adjustment. For instance, Mitic
and Rimer (2002) found that children in out-of-home care performed worse on standardized
assessments of reading and writing than their peers. Several other studies found that out-of-home
placement was disproportionately associated with poor performance on academic outcomes
(Colton et al., 1995; Sawyer & Dubowitz, 1994; Stein, 1997; Zima et al, 2000). Additionally,
prior studies have found that children in out-of-home placement were more likely to evidence
classroom behavior problems (Sawyer & Dubowitz, 1994; Stein, 1997; Zima et al, 2000).
Findings from this study also provide insight into the comorbidity of two of the social risks that
children with histories of out-of-home placement might experience. In this study, over a third of
the children with a history of out-of-home placement also had a history of substantiated child
maltreatment and over two-thirds had a history of homelessness. While other studies have
explored the overlap of these populations (Jonson-Reid, 2003; Zlotnick et al., 1999), few have
explored the prevalence of homelessness and maltreatment within a population of children with a
history of out-of-home placement. Prior studies have examined the co-occurrence of these risk
factors by calculating the prevalence of children with out-of-home placement experiences within
either the homeless population or the maltreated population (Jonson-Reid, 2003). This distinction
is critical because unlike homelessness and child maltreatment which are well documented risk
factors, out-of-home placement was developed to be an intervention for social risks such as child
maltreatment and homelessness — and not to be a risk factor unto itself. Wolkind and Rutter
(1973) note that while out-of-home placement has been ‘assumed’ to be the ‘harmful experience’,
“[w]ithout knowledge of the prior and subsequent family life experienced by children taken into
care, this is an unwarranted assumption”. This is similar to work that has been done in the area of
special education (Kauffman, 2005). Studies documented that children receiving special
education services did not perform as well as their peers not receiving these services. Yet, it was
not receipt of special education services that explained the disparity in outcomes, but rather the
constellation of risks/disabilities that led children to be placed in special education ([Kauffman,
2005]). This distinction brings to light the question of is out-of-home placement itself a risk, or
because it is an intervention for children who have experienced a myriad of social risks, is it a
proxy for the other risks that children have experienced prior to entering the foster care system?
The high co-occurrence between out-of-home placement with homelessness and child
maltreatment supported an examination of the mediating impact of child maltreatment and
homelessness on the relationship between out-of-home placement and educational well-being.
Diminished odds ratios were found for out-of-home placement — indicating that both
maltreatment and homelessness function as mediators for the relationship between out-of-home
placement and academic achievement and school adjustment.
Rutter (2000) and Waldfogel (2000a) emphasize the importance of examining the other risks
that children in out-of-home placement have been exposed. Although prior studies have
accounted for some of the variability in out-of-home placement experiences, they have not
accounted for how children's pre-placement experiences relate to their educational well-being.
This study's findings are also significant because a small number of studies have found that out-
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
957
Au
th
o
r's
pe
rs
on
al
co
py
of-home placement does not uniquely relate to children's academic outcomes when accounting
for other risks. For instance, studies conducted by Fox and Arcuri (1980) and Maluccio and Fein
(1985) found that when compared to children living in low-income households, children in outof-home placement were not at significantly greater risk for poor cognitive and academic
functioning. These findings suggest that out-of-home placement may be functioning through
other variables — such as maltreatment and homelessness.
This study provided the first careful examination of the extent to which two other social risk
factors, maltreatment and homelessness, function as mediating variables for the relationship
between out-of-home placement and children's educational well-being. It demonstrates the utility
of administrative data for understanding the relationship between children's early childhood
experiences and their educational well-being. This study also represents a collaborative effort
between a municipality and a university. However, study findings are qualified by the fact that the
data come from only one municipality and therefore are only generalizable to the demographic
composition of this northeast municipality. It will be important to replicate this study in other
municipalities to determine if the findings are generalizable.
Two additional areas of future research are an exploration of the timing of early childhood risks
and a more in depth exploration of the characteristics of out-of-home placement. The discovery of
child maltreatment and homelessness as mediating variables suggests that it is important to look at
the timing of these two events in relationship to out-of-home placement. A study by Park,
Metraux, Brodbar, and Culhane (2004) exploring the timing of homelessness and involvement in
the child welfare system found that 18% of children in the shelter system were involved with the
child welfare system within 5 years of exiting the shelter system and 6% received services prior to
entering the shelter system. A similar exploration of the timing of these events within early
childhood could help explain the nature of the mediating relationships of child maltreatment and
homelessness with out-of-home placement and provide additional information for informing
practice and policy. A second area of future research is the careful examination of how out-ofhome placement characteristics relate to educational well-being. Developing an understanding of
how specific out-of-home placement characteristics, such as age of entry and number of
disruptions, relate to educational well-being will provide key information for the formulation of
service plans for children in out-of-home placement.
This study generated two primary policy and practice implications. First, the high comorbidity
of out-of-home placement, child maltreatment and homelessness, as well as the risk for poor
educational outcomes associated with each of these social risks, underscores the importance of
collaboration within social service agencies and between social service agencies and public
school districts. Increased collaboration among social service agencies has the potential of
creating more comprehensive and integrated service delivery to children and their families
(Waldfogel, 2000b). Collaboration between child welfare and education professionals could
inform service planning and help enhance educational well-being for children experiencing these
early childhood risks. Altshuler (2003) outlined concrete benefits of routine meetings between
social service case managers and educators. For example, this contact could inform school
personnel of any changes in a child's placement that might affect the child's school performance
and provide an explanation for behavior change. Also, educators who are informed of a child's
service plan can support the family's participation with social services by linking participation to
positive educational outcomes.
Second, these findings underscore the importance of establishing more substantial ties
between child protective agencies and high quality early care and education programs. In this
study, over 75% of the children in the second grade cohort with a history of placement in out-of-
958
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
py
home care had their first placement before they were 5 years old. These data along with national
surveys (e.g., AFCARS, U.S. Department of Health and Human Services, 2003) emphasize the
importance of connecting this vulnerable group of children to beneficial early care and
educational experiences. Research supports well the protective value of participation in quality
early childhood programs; these programs have been found to promote early mastery of the
cognitive and social/emotional competencies that are necessary for early school success (NICHD,
2000a, 2002; National Research Council, 2000, 2001; Raver & Knitzer, 2002).
co
Acknowledgements
rs
on
al
This research project was supported in part by a grant from the William Penn Foundation and
federal grant #HD 046168, with funding from the Administration for Children and Families; the
Assistant Secretary for Planning and Evaluation; the US Department of Education: Office of
Special Education Programs, Institute for Education Sciences; and the National Institute of Child
Health and Human Development. This project was approved by the Kids Integrated Data System
(KIDS) Policy Review Committee. Special thanks go to the Julia Danzy, the Managing Director
of Social Services for the City of Philadelphia, Tom Clark, Director of the Office of Evaluation
and Research for the School District of Philadelphia, and Annabella Roig, the Co-Chair of the
KIDS Policy Review Committee for their collaboration and support. Correspondence concerning
this article should be addressed to the first author at The Penn Graduate School of Education,
University of Pennsylvania, 3700 Walnut Street, Philadelphia, PA 19104-6216.
pe
References
Au
th
o
r's
Aldgate, J., Colton, M., Ghate, D., & Heath, A. (1992). Educational attainment and stability in long-term foster care.
Children and Society, 6, 91−103.
Altshuler, S. J. (2003). From barriers to successful collaboration: Public schools and child welfare working together. Social
Work, 48, 52−63.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research:
Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173−1182.
Berrick, J. D., Needell, B., Barth, R. P., & Jonson-Reid, M. (1998). The tender years: Towards developmentally sensitive
child welfare services for very young children. New York: Oxford University Press.
Brazelton, B. T., & Greenspan, S. T. (2000). The irreducible needs of children. Cambridge, MA: Da Capo Press.
Canning, R. (1974). School experiences of foster children. Child Welfare, 53, 582−587.
Colton, M., & Heath, A. (1994). Attainment and behaviour of children in care and at home. Oxford Review of Education,
20, 317−327.
Colton, M., Heath, A., & Aldgate, J. (1995). Factors which influence the educational attainment of children in family foster
care. Community Alternatives: International Journal of Family Care, 7, 15−36.
Costello, E. J., & Angold, A. C. (2000). Developmental epidemiology: A framework for developmental psychopathology.
In A. J. Sameroff, M. Lewis, & S. M. Miller (Eds.), Handbook of developmental psychopathology (pp. 57−73). 2nd ed.
NY: Plenum Publishers.
CTB/McGraw–Hill. (1997). TerraNova. CA: Author.
CTB/McGraw–Hill. (2001). InView. CA: Author.
Dumaret, A. (1985). IQ, scholastic performance, and behavior of sibs raised in contrasting environments. Journal of Child
Psychology and Psychiatry, 26, 553−580.
Evans, L. (2004). Academic achievement of students in foster care: Impeded or improved? Psychology in the Schools, 41,
527−535.
Fantuzzo, J., Culhane, D. P., & Hadley, T. (2005). Kids Integrated Data System (KIDS).
Fanshel, D., & Shinn, E. (1978). Children in foster care: A longitudinal investigation. NY: Columbia University Press.
Fantuzzo, J. W., & Mohr, W. K. (2000). Prevalence and effects of child exposure to domestic violence. The Future of
Children, 9, 21−32 (Special Issue: Domestic violence and children).
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
959
Au
th
o
r's
pe
rs
on
al
co
py
Fox, M., & Arcuri, K. (1980). Cognitive and academic functioning in foster children. Child Welfare, 59, 491−496.
Goldstein, J., Solnit, A. J., Goldstein, S., & Freud, A. (1996). In the best interests of the child: The least detrimental
alternative. NY: Free Press.
Good, R. H., & Kaminski, R. A. (2002). Dynamic indicators of basic early literacy skills, 6th edition. OR: Institute for the
Development of Education Achievement.
Good, R. H., Wallin, J., Simmons, D. C., Kame'enui, E. J., & Kaminski, R. A. (2002). System-wide percentile ranks for
DIBELS benchmark assessment (Technical report no. 11). OR: University of Oregon.
Heath, A. F., Colton, M. J., & Aldgate, J. (1994). Failure to escape: A longitudinal study of foster children's educational
attainment. British Journal of Social Work, 24, 241−260.
Horwitz, S. M., Simms, M. D., & Farrington, R. (1994). Impact of developmental problems on young children's exits from
foster care. Journal of Developmental and Behavioral Pediatrics, 15, 105−110.
Huffman, L.C., Mehlinger, S.L., & Kerivan, A.S. (2000). Risk factors for academic and behavioral problems at the
beginning of school. In Off to a good start: Research on the risk factors for early school problems and selected federal
policies affecting children’s social and emotional development and their readiness for school. National Research
Council. (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press.
Jonson-Reid, M. (2003). Foster care and future risk of maltreatment. Children and Youth Services Review, 25, 271−294.
Kauffman, J. M. (2005). Point of view: Waving to Ray Charles — Missing the meaning of disabilities. Phi Delta Kappan,
86, 520.
Klee, L., Kronstadt, D., & Zlotnick, C. (1997). Foster care's youngest: A preliminary report. American Journal of
Orthopsychiatry, 67, 290−299.
Leslie, L. K., Gordon, J. N., Ganger, W., & Gist, K. (2002). Developmental delay in young children in child welfare by
initial placement type. Infant Mental Health Journal, 23, 496−516.
Maluccio, A. N., & Fein, E. (1985). Growing up in foster care. Children and Youth Services Review, 7, 123−134.
Masten, A. S., & Wright, M. O. (1998). Cumulative risk and protection models of child maltreatment. Journal of
Aggression Maltreatment, & Trauma, 2, 7−30.
McCue-Horwitz, S., Balestracci, K. M. B., & Simms, M. D. (2001). Foster care improves children's functioning. Archives
of Pediatrics & Adolescent Medicine, 155, 1255−1260.
McDermott, P. A., Green, L. F., Francis, J. M., & Stott, D. H. (1999). Learning Behaviors Scale. Philadelphia: Edumetric
and Clinical Science.
McWayne, C. M., Fantuzzo, J. W., & McDermott, P. A. (2004). Preschool competency in context: An investigation of the
unique contribution of child competencies to early academic success. Developmental Psychology, 40, 633−645.
Microsoft SQL Server. (2005).
Mitic, W., & Rimer, M. L. (2002). The educational attainment of children in care in British Columbia. Child and Youth
Care Forum, 31, 397−414.
Nash, J. K., & Bowen, G. L. (2002). Defining and estimating risk and protection: An illustration from the School Success
Profile. Child & Adolescent Social Work Journal, 19(3), 247−261.
National Center for Children in Poverty. (2001). Improving the odds for the healthy development of children in foster care.
New York, NY: National Center for Children in Poverty.
National Institute of Child Health and Human Development. (2000a). The relation of child care to cognitive and language
development. Child Development, 71, 960−980.
National Institute of Child Health and Human Development. (2000b). Teaching children to read: An evidence based
assessment of the scientific research literature on reading and its implication for reading instruction. NIH Publication,
Vol. 00-4769. Washington, D.C.: U.S. Government Printing Office.
National Institute of Child Health and Human Development. (2002). Early child care and children's development prior to
school entry: Results from the NICHD study of early child care. American Educational Research Journal, 39, 133−164.
National Research Council. (1998). Preventing reading difficulties in young children. Washington, DC: National Academy
Press.
National Research Council. (2000). From neurons to neighborhoods: The science of early childhood development.
Washington, DC: National Academy Press.
National Research Council. (2001). Eager to learn: Educating our preschoolers. Washington, DC: National Academy Press.
Park, M. J., Metraux, S., Brodbar, G., & Culhane, D. P. (2004). Child welfare involvement among children in homeless
families. Child Welfare, 83, 423−436.
Pearson Reading Group. (2003). Developmental Reading Assessment (DRA).
Reams, R. (1999). Children birth to three entering the state's custody. Infant Mental Health Journal, 20, 166−174.
Raver, C.C., & Knitzer, J. (2002). Ready to enter: What research tells policy makers about strategies to promote social and
emotional school readiness among three and four year old children. Paper series in the “Promoting the emotional well-
960
J. Fantuzzo, S. Perlman / Children and Youth Services Review 29 (2007) 941–960
Au
th
o
r's
pe
rs
on
al
co
py
being of children and families”. National Center for Children in Poverty. Mailman School of Public Health, Columbia
University.
Rutter, M. (2000). Children in substitute care: Some conceptual considerations and research implications. Children and
Youth Services Review, 22, 685−703.
Sawyer, R. J., & Dubowitz, H. (1994). School performance of children in kinship care. Child Abuse and Neglect, 18,
587−597.
Scott, K. G., Mason, C. A., & Chapman, D. A. (1999). The use of epidemiological methodology as a means of influencing
public policy. Child Development, 70, 1263−1272.
Silver, J., DiLorenzo, P., Zukoski, M., Ross, P. E., Amster, B. J., & Schlegel, D. (1999). Starting young: Improving the
health and developmental outcomes of infants and toddlers in the child welfare system. Child Welfare, 78, 148−165.
Stein, E. (1997). Teachers' assessments of children in foster care. Developmental Disabilities Bulletin, 25, 1−17.
Stott, D. H., McDermott, P. A., Green, L. F., & Francis, J. M. (1988). Study of children's learning behaviors. San Antonio:
The Psychological Corporation.
U.S. Department of Education. (2004). Four Pillars of NCLB. http://www.ed.gov/nclb/overview/intro/4pillars.html.
Retrieved March 19, 2004, from.
U.S. Department of Health and Human Services. (2003). The AFCARS report. http://www.acf.hhs.gov/programs/cb/
publications/afcars/report8.pdf. Retrieved on July 18, 2003, from.
Waldfogel, J. (2000a). Child welfare research: How adequate are the data? Children and Youth Services Review, 22,
705−741.
Waldfogel, J. (2000b). Reforming child protective services. Child Welfare, 79, 43−57.
Weiss, A. D., & Fantuzzo, J. W. (2001). Multivariate impact of health and caretaking risk factors on the school adjustment
of first graders. Journal of Community Psychology, 29, 141−160.
Whiting-Blome, W. (1997). What happens to foster kids: Educational experiences of a random sample of foster care youth
and a matched group of non-foster care youth. Child & Adolescent Social Work Journal, 14, 41−53.
Wolkind, S., & Rutter, M. (1973). Children who have been “in care” — An epidemiological study. Journal of Child
Psychology and Psychiatry, 14, 97−105.
Zima, B., Bussing, R., Freeman, S., Yang, X., Belin, T., & Forness, S. (2000). Behavior problems, academic skill delays
and school failure among school-aged children in foster care: Their relationship to placement characteristics. Journal
of Child and Family Studies, 9, 87−103.
Zlotnick, C., Robertson, M. J., & Wright, M. A. (1999). The impact of childhood foster care and other out-of-home
placement on homeless women and their children. Child Abuse and Neglect, 23, 1057−1068.