Advances in Applied Sociology
2013. Vol.3, No.8, 320-328
Published Online December 2013 in SciRes (http://www.scirp.org/journal/aasoci) http://dx.doi.org/10.4236/aasoci.2013.38041
Open Access
320
Keeping the White Family Together: Racial Disparities in the
Out-of-Home Placements of Maltreated Children
Angela M. Kaufman
Bowling Green State University, Bowling Green, USA
Email: kaufmaa@bgsu.edu
Received September 24th, 2013; revised October 24th, 2013; accepted October 31st, 2013
Copyright © 2013 Angela M. Kaufman. This is an open access article distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the
original work is properly cited.
The likelihood of being removed from the home following a case of maltreatment is much higher for
black youth than for whites. Two explanations exist in the literature. The first, black children experience
more serious forms of maltreatment and have fewer resources to remedy the maltreatment situation than
do whites. The second, there is an underlying racial bias within the child welfare system. The present
study examines 789 dependency cases from child welfare services in a large urban county in the North-
west United States. Using multiple logistic regression models, it examines whether race has an effect on
child placement within the child welfare system, and whether the factors influencing placement are the
same for white and black youth. Findings illustrate a racial disparity in out-of-home placements support-
ing both of the competing explanations in the current literature. Overall, the present study finds that two
separate processes seem to be at play in the placement decisions of maltreated youth, and concludes with
possible explanations for this differential treatment.
Keywords: Child Welfare; Child Maltreatment; Foster Care; Child Service Agencies; Out-of-Home
Placement, Child Abuse; Racial Differences in Child Placement
Introduction
More than 3.5 million children received Child Protective
Service (CPS) investigations in 2007, with an estimated
794,000 children found to have substantiated cases of mal-
treatment (US Dept., 2009). While this translates into 10.6 per
1000 children experiencing maltreatment in the population
overall, this rate varies considerably across racial-ethnic groups
(US Dept., 2009). More importantly, these same disparities can
be seen when looking at the intervention decisions concerning
the maltreated child, where studies show that children of mi-
nority groups, particularly the children of black families, are
placed in foster care at higher rates than children from white
families (Needell, Magruder, & Putnam-Hornstein, 2007; Ri-
vaux et al., 2008). Examining decision making within CPS is
important, as removing maltreated children from their homes
can have lifelong consequences on their ability to form bonds
and attachments with others. Out-of-home placement for mal-
treated children may also increase their risk for later juvenile
and adult criminal offending, depending on the circumstances
of their removal and placement (Ryan, Testa, & Zhai, 2008).
Examining placement decisions in the earlier stages of the child
welfare process, where caseworkers and judges have more dis-
cretion, will also allow for the understanding of where in the
process unwarranted disparities might develop.
Background
Two Explanations for Racial Disparities
Much prior research has analyzed the causes of child mal-
treatment (Barth, 2009; Barth, Wildfire, & Green, 2006; Phil-
lips, Burns, Wagner, & Barth, 2004; Polansky, Chalmer, But-
tenweiser, & Williams, 1981), but much less has examined the
placement decisions associated with maltreatment. Especially
neglected until recently is whether race and ethnicity play a role
within child welfare; and findings are inconsistent among the
few studies that do look at this issue. Some argue the overrep-
resentation of blacks within CPS can be accounted for by dif-
ferences in poverty and other forms of structural disadvantage;
while others state there may be additional factors at play, per-
haps racial bias, which contributes to the greater representation
of black children in the system (Barth et al., 2006; Lindsey,
1994; Phillips et al., 2004; Pinderhughes, 1991).
One explanation for the greater likelihood of out-of-home
placement among black children is that disadvantaged locations
of racial minority groups put them at higher risk for maltreat-
ment, and makes the evidence of maltreatment more visible to
CPS agencies (Drake, Lee, & Jonson-Reid, 2009; Knott & Do-
novan, 2010). Characteristics contributing to higher levels of
maltreatment and greater intervention include lower socioeco-
nomic status, greater family instability, and parental health
problems. These same characteristics are often viewed as more
difficult to remedy through both community and official inter-
ventions (Brown, 2008). However, an alternative argument is
that racial disparities exist in part because of an underlying bias
within the system (Hampton & Newberger, 1985; Hill, 2004;
Knott & Donovan, 2010; Miller & Gaston, 2003; Osterling,
D’Andrade, & Austin, 2008). This research finds that racial
minorities are at no greater risk of maltreatment than are whites,
even when disadvantage is taken into account. If these findings
A. M. KAUFMAN
represent the true nature of racial disparity in child welfare, it is
possible that part of this disparity is due to judgments based on
race.
The Race-Poverty Link
The association between race and poverty is one of long-
recognized significance. In 2007, 8.6 percent of non-Hispanic
whites were declared impoverished according to national pov-
erty thresholds, while 24.7 percent of blacks met such standards
(National, 2007). This association, in turn, may have negative
repercussions for racial-ethnic minorities in child welfare. For
instance, according to Lindsey’s (1994) analysis of national
survey data, parent’s income level was the major determinant in
a child’s removal from his or her family. While this analysis did
not focus on race specifically, economic inequalities that dis-
advantage minorities prevent the informal rectification of child
maltreatment once reported. Specifically, CPS agencies in 33
states in the US reported that high community poverty rates
may increase the proportion of black children entering foster
care compared to whites, who are much less likely to live in
impoverished neighborhoods (Brown, 2008). This is because
residing in poor communities oftentimes limits access to the
kinds of support and services needed to both prevent and rectify
child maltreatment (Rivaux et al., 2008; Schuck, 2005). Neces-
sary support and services include affordable and adequate
housing, substance abuse treatment, and family counseling
(Brown, 2008). Mental health services are particularly impor-
tant for black youth, who have been found to demonstrate the
greatest level of need compared to both white and Hispanic
youth, but are the least likely to receive such services (Rawal,
Romansky, Jenuwine, & Lyons, 2004). Therefore, without the
ability to remedy the situation in the allotted timeframe be-
tween case reporting and appearing before the dependency
judge, an out-of-home placement may be seen as the only fea-
sible alternative in alleviating the maltreated child’s suffering.
However, prior research supporting a race-poverty explanation
for racial disparities in the child welfare system still finds that
socioeconomic status cannot fully explain the race gap. Black
children have been found to experience upwards of a 77 percent
increased likelihood of home removal after such variables as
household income are taken into account (Rivaux et al., 2008).
The Importance of Familial Characteristics
In addition to poverty, other characteristics may contribute to
the racial disparity in out-of-home placements. One such char-
acteristic is family structure. Poverty rates are highest for fami-
lies headed by single women, particularly if they are black or
Hispanic (National, 2007). Related, both Hispanic and non-
Hispanic black women have more unmarried childbirths than
do whites (Martin et al., 2009). This may be especially relevant
in cases of neglect, which instead of being malicious is often
the unfortunate consequence of lacking the resources necessary
to provide a healthy and safe environment for children. Simi-
larly, the poor, and often racial minority, are likely to have co-
occurring problems of substance abuse, arrest or incarceration,
mental health problems, chronic illness, and low education
(Barth et al., 2006; Phillips et al., 2004; Polansky et al., 1981).
These problems increase the likelihood that such families will
become involved with CPS agencies through avenues other
than the maltreatment itself, supporting the possible legitimacy
of racial disproportionality in CPS agencies.
Understanding Disparity within the Child Welfare
System as Resulti ng fr om Bias
The belief that biases and discrimination are behind racial
disparities in CPS agency involvement is supported by several
national studies suggesting there are no racial-ethnic differ-
ences in the occurrence of child maltreatment (Kirk & Griffith,
2008; Sedlack & Broadhurt, 1996). Prior studies also indicate
that black families are held to different standards based on their
perceived dangerousness and threat to mainstream society (Al-
bonetti, 1991; Bridges & Steen, 1998; Hill, Harris, & Miller,
1985; Sampson, 1986; Tonry, 1995). Thus, differences in out-
comes between black and white children in CPS may be largely
due to the perception that blacks constitute a more dangerous
group, even when their behaviors are similar to that of whites
(Albonetti, 1991). While empirical literature on overt racism
within the child welfare system is scarce, evidence from case
review studies often support racial bias in child maltreatment
reporting (Hampton & Newberger, 1985; Lane, Rubin, Mon-
teith, & Christian, 2002). Related, some researchers and practi-
tioners state that a universal set of standards is oftentimes used
to evaluate families, without taking into account social and
cultural diversity. As such, characteristics of poverty or single
parenthood may be seen as deviant in the system if, for exam-
ple, two-parent middle-class households are held as the stan-
dard of a well-functioning family (Billingsley & Giovannoni,
1972; Hill, 2004; Miller & Gaston, 2003). This phenomenon is
especially evident in less serious reports, where case workers
and judges have greater discretion, and are thus more suscepti-
ble to the practice of differential response (Osterling et al., 2008;
Rivaux et al., 2008). Explanations for this differential response
have been attributed to a variety of factors, including: black
families’ abuse and neglect being seen as less remediable than
white families’, black families being held to unattainable white
middle-class standards, or black families suffering from the
devaluation of their culture and family functioning (Billingsley
& Giovannoni, 1972; Hill, 2004; Knott & Donovan, 2010).
Study Significance
Two main research questions guide this study. First, does a
maltreated child’s race affect the likelihood of receiving an
out-of-home placement? Second, do different factors matter for
black and white children within the decision making process?
Based on prior research, I hypothesize black children will be
more likely to have an out-of-home placement than their white
counterparts. However, I hypothesize that some of the race gap
is explained by risky familial characteristics such as low socio-
economic status, substance abuse, and mental illness that may
be more common among blacks than whites. I also hypothesize
that these risk factors exert a stronger effect among blacks than
whites, as these characteristics may serve to reinforce beliefs
about the perceived inability of black families to care for their
children. They may also illustrate the possibility that the thres-
hold when identifying these factors as problematic is lower for
black families (Sampson & Laub, 1993; Tonry, 1995). Finally,
I hypothesize that characteristics specific to maltreatment, such
as the type of abuse, will matter more for white children than
for black children. By controlling for characteristics associated
with both racial minority status and out-of-home placement, I
hope to better understand the extent to which racially-disparate
outcomes in CPS are unwarranted. In doing so, this study will
enable CPS and other law enforcement officials to tailor their
Open Access 321
A. M. KAUFMAN
response and treatment of minority groups accordingly in han-
dling child maltreatment cases.
Data and Measures
The dataset I use to conduct this study is “Childhood Vic-
timization and Delinquency, Adult Criminality, and Violent
Behavior in a Large Urban County in the Northwest United
States” from the Inter-University Consortium for Political and
Social Research (ICPSR) online database. The dataset consists
of children who were age birth to 11 years between 1980 and
1984, and were born in a large urban county in the Northwest
United States. Data was obtained from administrative records;
including birth records and county court house dependency
records, as well as from the US Census Bureau for the socio-
economic variables. Potential respondents were selected based
on whether the child was born in the state, their maltreatment
case stayed within the county during the study and they were
made a dependent of the state within that county, their depend-
ency record was available for data documentation, and the child
was still alive, i.e. did not die as a result of maltreatment or
other causes throughout the study. After these rules were ap-
plied, 877 dependency petitions were included.
Dependent Variable: Child placement serves as the outcome
variable in this analysis and is divided into several categories:
remained with parent or guardian, group home, adopted, kin
care, foster care and aged out of the system. In this analysis,
placement is dichotomized with an in-home placement coded as
0 and including only those children ordered to remain with their
parents or guardian, while out-of-home placement is coded as 1,
including all other categories, except those cases in which the
subject aged out of the system. These individuals are not eligi-
ble for placement and are deleted from the sample.
Independent Variable: The main independent variable is the
racial identification of the maltreated child. Although the origi-
nal coding of this variable categorizes children as Native Amer-
ican, African American, American Asian, American Pacific
Islander, Caucasian, and Other classification, due to small sam-
ple sizes in some of these categories, the race variable is re-
coded into a white (coded as 0 for the reference category) and
black (coded as 1) dichotomous variable, while all other racial-
ethnic categories, 66 cases representing less than eight percent
of the sample, are dropped from the analyses.
Relevant Placement Factors: These include the characteris-
tics of the maltreatment, as well as other disadvantageous fa-
milial characteristics that could be detrimental to the child’s
well-being, regardless of whether the maltreatment alone war-
rants home removal. These are important to take into account,
as prior research claims that it is both abuse characteristics and
other co-occurring issues that account for the higher representa-
tion of black children within the system (Barth, 2009; Burns et
al., 2004; Courtney, McMurtry, & Zinn, 2004). Maltreatment
Type: The measure of maltreatment type differentiates between
four major types of maltreatment: physical abuse, sexual abuse,
emotional abuse, and neglect, as well as a fifth category of
multiple abuse types which encompasses those children who
have experienced any combination of two or more types of the
major maltreatment categories. These five categories of mal-
treatment will be dichotomized into a 0, 1 coding, indicating
which type of maltreatment each subject experienced. Sibling
Victimization: Whether another child is being victimized in the
home demonstrates the overall safety of the home environment,
as well as the likelihood that the subject may suffer further
maltreatment. Sibling victimization is dichotomized into a 0, 1
coding, signifying whether a sibling of the subject has also
suffered from abuse or neglect. AFDC: Poverty has been cited
as a main consideration in removing a child from the home
(Barth, 2009; Lindsey, 1994; Rivaux et al., 2008; Schuck, 2005)
and is highly correlated with racial-ethnic status (National,
2007). As such, in this analysis, Aid to Families with Depend-
ent Children (AFDC) payments will be used as a proxy for the
measurement of parental socioeconomic status. The variable
AFDC reports the percent of families receiving such payments
in the subject’s birth census tract based on the 1980 census, and
is a continuous measurement at the census tract level. Parental
Health: Parental mental illness and substance abuse are impor-
tant characteristics to take into account as they contribute to the
possible instability of the home environment, and have been
found in prior research to be correlated with racial minority
status (Barth et al., 2006; Phillips et al., 2004). Both of these
measures are dichotomized into a 0, 1 coding, signifying whe-
ther the parent(s) displays such characteristics when the mal-
treatment case comes before the dependency judge at the initial
court hearing. Behavioral Problems: An estimated 19 percent
of children in the National Survey of Child and Adolescent
Well-Being (NCSAW) entered out-of-home placement, even
though they did not have obviously unfit parents. This was
suggested to have occurred due to the multitude of emotional
and behavioral difficulties the child was dealing with (Barth et
al., 2006). Child behavioral problems are dichotomized into a 0,
1 coding, signifying whether the child displays such charac-
teristics when the maltreatment case comes before the depend-
ency judge at the initial court hearing.
Demographic Controls: Both the gender and age of the child
are controlled for in this analysis. Gender is highly correlated
with some types of abuse, specifically sexual abuse, which
females experience at a much higher rate than males (Widom &
Maxfield, 2001). In regard to age, NCANDS data illustrate
(Wulczyn, 2009) that younger children may be at a higher risk
for both the actual occurrence of maltreatment, as well as it
being reported. This finding can be extrapolated to the place-
ment of the child, in that younger children may be perceived as
having a higher risk for subsequent victimization. Gender will
be dichotomized, with male coded as 0 for the reference cate-
gory, while age is measured using a continuous variable.
Analytic Strategy
The original data sample is restricted in three ways for this
study’s analyses. First, when looking at the selected variables
for this analysis, the data completeness report shows that only
one of the 877 treatment cases is missing. Therefore, this one
individual will be dropped from the analysis. Second, those
subjects falling into racial categories other than white and black
are excluded. Finally, those subjects that aged out of the system
between their maltreatment report and appearing before the
judge at the initial dependency hearing are dropped from the
analyses, as they are not eligible for a placement decision.
These restrictions result in a total sample size of n = 789, with
598 white and 191 black children. With these restrictions in
place, I will first examine descriptive statistics focusing on
differences between black and white children on the dependent
variable and important risk factors. Second, I will use logistic
regression to examine the differences for placement outcomes
Open Access.
322
A. M. KAUFMAN
Open Access 323
of the maltreated children. In a series of nested models, I will
analyze what relevant risk and demographic factors are associ-
ated with placement decisions and the extent to which they
account for any black-white differences. In the first set of ana-
lyses, a total of five nested logistic regression models will be
used for the full sample of maltreated children, regressing the
placement decision on the key independent variable, race, de-
mographic controls for the child, and other relevant risk factors
related to both the child and parents generally, as well as the
maltreatment situation specifically. In the second set of analy-
ses, the full model will then be partitioned into two separate
models by race. Z-tests for coefficient differences will be per-
formed to examine if the risk factors have different effects for
black and white children within the child welfare process.
Results
Descriptive Statistics
In Table 1, I provide descriptive statistics for all variables in
the analysis. I show the characteristics of the total sample (n =
789), and for the separate white (n = 598) and black (n = 191)
samples. Also, I provide results of 2 and t-tests to denote
characteristics where white and black youth differ. Most nota-
bly is the significant difference in the dependent variable,
placement. About 74 percent of black children are ordered out
of the home, compared to only about 46 percent of white child-
ren. However, black and white children vary little on many
relevant placement characteristics, including gender, age, pa-
rental substance abuse and parental mental illness, childhood
behavioral problems and physical abuse as the maltreatment
type. Risk factors that are more common among black children
include AFDC (15.0% versus 7.0%), and maltreatment types of
neglect and emotional abuse (39.3% versus 30.4%, and 16.8%
versus 6.9%, respectively). Relevant risk factors that are more
prevalent among white children, on the other hand, include
sibling victimization (17.4% versus 1.6%), and sexual and mul-
tiple abuse as the documented forms of maltreatment (10.9%
versus 4.7% and 42.8% versus 29.8%, respectively).
Multiple Regression Models-Ordered Placement for
All Maltreated Children
In the next part of the analysis, I examine whether the racial
disparity in the likelihood of receiving an out-of-home place-
ment can be explained by those risk factors that black youth are
more likely to experience than their white counterparts. Table 2
examines the relationship between race and placement for the
full sample of maltreated children in a series of nested models,
controlling for demographic characteristics, family risk factors
and maltreatment characteristics.
In Model 1, placement is regressed only on the main inde-
pendent variable, race. The odds of receiving an out-of-home
placement are 3.335 times higher for black children than for
white. In Model 2, the age and gender of the maltreated child
are entered into the equation. While the effect of age is not
statistically significant, the effect of gender is. The odds of
being ordered out of the home are 32% lower for girls than for
boys. However, controlling for age and gender do not appear to
substantially reduce the higher risk of black children being
removed from the home compared to white children. In Model
3, I include the AFDC measure to examine whether socioeco-
nomic status accounts for any of the racial disparity in place-
ment. Consistent with findings in prior research, the effect of
AFDC is statistically significant. For every percentage increase
Table 1.
Descriptive statistics for white and black maltreated children.
Measure Total Mean White Mean Black Mean Difference Tests
Out-of-home placement 52.6% 45.8% 73.8% 45.5290***
Female 52.9% 54.5% 47.6% 2.7427
Age 7.00 6.98 7.07 0.4074
AFDC 8.92 6.99 14.98 13.5239***
Parental substance abuse 24.8% 24.1% 27.2% 0.7668
Parental mental illness 16.9% 16.4% 18.3% 0.3874
Childhood behavioral problems 14.2% 15.2% 11.0% 2.1192
Sibling victimization 13.6% 17.4% 1.6% 30.9094***
Neglect 32.6% 30.4% 39.3% 5.1417*
Emotional abuse 9.3% 6.9% 16.8% 16.8907***
Physical abuse 9.1% 9.0% 9.4% 0.0271
Sexual abuse 9.4% 10.9% 4.7% 6.4578*
Multiple abuse types 39.7% 42.8% 29.8% 10.1695***
N 789 598 191
Note: *p < 0.05, **p < 0.01, ***p < 0.001.
A. M. KAUFMAN
Table 2.
Ordered placement for all maltreated children.
Regressor Model 1 Odds Model 2 Odds Model 3 Odds Model 4 Odds Model 5 Odds
Intercept 0.168*
(0.082) 0.055
(0.216) 0.220
(0.225) 0.526*
(0.240) 0.132
(0.278)
Black 1.204***
(0.184) 3.335 1.187***
(0.185) 3.279 0.962***
(0.201) 2.616 1.037***
(0.205) 2.822 0.877***
(0.212) 2.404
Female
0.386**
(0.148) 0.680 0.416**
(0.149) 0.659 0.179
(0.158) 0.836 0.089
(0.163) 0.915
Age
0.014
(0.027) 1.1014 0.010
(0.027) 1.010 0.027
(0.029) 0.974 0.022
(0.029) 0.978
AFDC
0.030*
(0.011) 1.031 0.035**
(0.011) 1.035 0.032**
(0.012) 1.033
Parental substance
abuse
0.382*
(0.178) 1.465 0.315
(0.181) 1.370
Parental mental illness 0.497*
(0.209) 1.645 0.310
(0.214) 1.363
Prior childhood
behavioral problems
1.530***
(0.258) 4.618 1.416***
(0.261) 4.120
Sibling victimization
0.591*
(0.242) 0.554
Emotional abuse
0.259
(0.305) 1.295
Physical abuse
0.628*
(0.291) 0.534
Sexual abuse
0.860**
(0.309) 0.423
Multiple abuse types 0.392*
(0.278) 0.675
χ2 47.22*** 54.18** 62.07*** 110.74*** 132.60***
Psuedo R2 0.0433 0.0496 0.0569 0.1014 0.1215
Note: *p < 0.05, **p < 0.01, ***p < 0.001.
in the families receiving AFDC payments in the subject’s cen-
sus tract, the odds of out-of-home placement increase by about
3 percent. Furthermore, the inclusion of this variable substan-
tially reduces the disparity between black and white children in
the odds of being ordered out of the home, from 3.28 in Model
2 to 2.616 in Model 3, a reduction of about 20 percent. How-
ever, the effect of race is still statistically significant.
Model 4 adds familial characteristics that may further con-
tribute to disparity, as they illustrate relevant risk separate from
maltreatment that may be detrimental to the child’s well-being.
All three characteristics, parental substance abuse, parental
mental illness, and prior childhood behavioral problems, are
significantly related to being removed from the home, with
increased likelihoods of 46.5%, 64.5%, and 361.8%, respec-
tively. The inclusion of this additional block of variables also
reduces the effect of female to insignificance. Entering the va-
riables individually shows that it is the addition of prior child-
hood behavioral problems that has this effect, as males are
more likely to exhibit such behaviors in comparison to females.
Including these three variables also slightly increases the racial
disparity in out-of-home placement. Whereas in Model 3, the
odds of being ordered out of the home were 2.616 higher for
black children, net of other factors, this likelihood increases to
2.822 in Model 4. Again, this is due to the inclusion of prior
childhood behavioral problems. Adding the variables individu-
ally illustrates that parental mental illness and parental sub-
stance abuse affect the coefficient for black by less than 1 per-
cent, whereas prior childhood behavioral problems lead to an
increase of approximately 8 percent in the race coefficient in
Model 4. Put simply, despite having fewer documented behav-
ioral problems, black children are still more likely to be re-
moved from the home than whites, net of other factors.
The final model (Model 5) illustrates that the type of mal-
treatment the child experiences is a significant predictor of out-
of-home placement, with children experiencing physical abuse,
sexual abuse, and multiple abuse types having significantly
different odds of being removed from the home when compared
to neglect. Interestingly, these three types of maltreatment actu-
ally decrease the odds of out-of-home placement, net of other
factors. There is no significant difference for those children ex-
periencing emotional abuse compared to neglect. Counterintui-
tively, those children with a sibling who has also been mal-
treated are about 45% less likely to be removed from the home,
net of all factors in the model.
Open Access.
324
A. M. KAUFMAN
The significant reduction in the odds of being removed from
the home for those children who experience physical, sexual
and multiple abuse types, illustrates that the child welfare sys-
tem may not aim to remove children from an abusive home, but
rather to remove children from homes in which there is a gen-
eral inability to care. This may be due to the perception that
abuse is an infrequent, rectifiable form of maltreatment, where-
as neglect implies continuous and accumulating negative cir-
cumstances that cannot be remedied through the provision of
in-home services. This would also explain why there are no
significant differences in the odds of out-of-home placement
between those children who experience neglect and those who
are emotionally abused. While emotional abuse encompasses a
wide range of behaviors, bivariate analyses indicate that the
most common form of emotional abuse experienced by children
removed from the home is being abandoned for 24 hours or
more. Therefore, emotional abuse, at least according to the
Maltreatment Coding Scheme used for the present data collec-
tion, is actually a very specific form of what is traditionally
thought of as neglect, and thus also represents a more enduring
form of maltreatment in comparison to sexual, physical and
multiple abuse types. This possible explanation is also consis-
tent with national data. For instance, in fiscal year 2007, of the
1760 children who died due to child abuse or neglect, 34.1
percent of child fatalities were attributed to neglect only, this
percentage not including those children who suffered from
multiple forms of maltreatment which also included neglect,
and thus may far underestimate the true severity of maltreat-
ment in regard to neglect (US Dept., 2009). Moreover, of the
approximately 265,000 children who were removed from their
homes during the same fiscal year, 69.2 percent were victims of
neglect, whereas 8.6 percent suffered from physical abuse, 14.2
percent from multiple forms of maltreatment, and only 3.2 per-
cent from sexual abuse (US Dept., 2009).
A second possible explanation for the finding that sexual,
physical and multiple maltreatment types reduce the odds of
out-of-home placement in comparison to neglect is that perpe-
trators of neglect are more likely to be the parents of child ne-
glect victims. Therefore, out-of-home placement is the most
reasonable solution for children who are neglected, as their
offenders usually reside in the same household. For other types
of maltreatment, though, parents may not be the perpetrators,
and thus the child can be removed from the dangers of such
abuse without being removed from the home. This second ex-
planation is supported by 2007 National Child Abuse and Neg-
lect Data System (NCANDS) data, where the percentage of
perpetrators of sexual abuse was highest among friends or
neighbors at 57.7 percent and child daycare providers at 23.9
percent, compared to only 2.4 percent of such perpetrators be-
ing parents. For physical abuse, child daycare providers consti-
tuted 14.1 percent of perpetrators, and friends and neighbors
constituted another 14.4 percent, compared to only 9.7 percent
whom were parents. For neglect cases, on the other hand, 66.1
percent of perpetrators were the parents of the victim (US Dept.,
2009). However, the finding that having a sibling who has also
been victimized in some form reduces the odds of out-of-home
placement, net of all factors in the model, is still unclear.
The inclusion of maltreatment specific variables reduces the
racial disparity in out-of-home placement from black children
having 2.822 to 2.404 higher odds of being removed from the
home compared to white children from Model 4. Additionally,
these measures reduce the effects of both parental substance
abuse and parental mental illness to insignificance, indicating a
possible mediation effect. For instance, parents experiencing
substance abuse and mental health issues may unintentionally
emotionally abuse or neglect their children as a result of such
illnesses, and these two types of maltreatment are the forms
most likely to result in out-of-home placements among mal-
treated youth. Importantly, while the effect of race remains
statistically significant in all models, about 28% of the racial
disparity is explained from the first to the final model in the
series of regressions.
Ordered Placement by Race
In the analyses that follow (not shown here due to limited
significant findings), the full sample is partitioned by race to
see what, if any, factors matter differently for white and black
maltreated children in predicting the likelihood of out-of-home
placement. Relevant risk and other factors are entered into the
regression in the same order for the partitioned models as was
the case for the full sample, and differences in coefficients be-
tween the two models are compared using z-tests. It is also
important to note that due to the small size of the black sample,
there is a reduction in statistical power in the partitioned models,
rendering it more difficult to find significant effects within the
black model, as well as significant differences between the
black and white models. For the model including only black
children, AFDC and emotional abuse, in comparison to neglect,
are the only variables which aid in the prediction of home re-
moval. For the model including only white children, parental
mental illness, prior child behavioral problems, sibling victimi-
zation, and sexual abuse are the significant predictor variables.
However, z-tests indicate that the only differences across race
are for parental mental illness and emotional abuse as the mal-
treatment type. For white children, parental mental illness in-
creases the odds of being removed from the home by about 81
percent, while for black children there is no significant effect.
Emotional abuse, on the other hand, increases the odds of out-
of-home placement by a factor of 6.213 for black children,
while there is no significant effect of emotional abuse among
white children.
While examining these effects separately may not provide
much understanding in regard to racial disparity in CPS, two
propositions can be made upon examining these effects simul-
taneously, based on prior research in the juvenile justice and
child maltreatment literatures. Going back to the bivariate re-
sults in Table 1, there is no significant difference in the number
of parents suffering from mental illness among white and black
children. Additionally, parental mental illness and emotional
abuse are significantly positively correlated. Yet parental men-
tal illness is a significant predictor of out-of-home placement
only for whites, while emotional abuse is significant only for
blacks.
One possible explanation can be understood through the
work of Bridges and Steen (1998), where they find that while
the offenses of minority youth are often seen as individual fail-
ings in the form of negative attitudinal and personality traits,
these same offenses are often portrayed as being a result of
negative environmental factors for white adolescents. Extrapo-
lating these findings to the arena of child maltreatment and
adult offenders, then, allows for the hypothesis that among
parents who maltreat their children, black parents will be seen
as personally responsible for the abuse and neglect their child-
Open Access 325
A. M. KAUFMAN
ren suffer, while factors beyond the individuals’ control will be
used to remove accountability from white parents. Thus, white
parents victimize children as a result of mental illnesses, medi-
cal conditions beyond their control which cause them to be
neglectful, aggressive, impulsive, and so on. Black parents, on
the other hand, despite having a diagnosed mental illness still
choose to maltreat their children, according to this perspective.
Whether this is because they have chosen not to seek treatment
for their illness, or because it is believed the maltreatment
would occur regardless of the presence of illness, black parents
are seen as solely responsible for maltreatment. A second pos-
sible explanation, related to the first, is the effect of socioeco-
nomic status in the acceptance of a mentally ill label. Bivariate
analyses indicate that there is a statistically significant differ-
ence in approximate socioeconomic status between white and
black maltreated children. Moreover, the partitioned models
illustrate that while the percent of families receiving AFDC
payments within the subject’s census tract is a significant pre-
dictor in being removed from the home for black children, there
is no significant effect for white children. As illustrated by
Brown (2008), black children may be at an increased risk for
out-of-home placement due to living in impoverished neigh-
borhoods where access to the kinds of support and services
needed to both prevent and rectify child maltreatment is limited.
One of these suggested services is mental health and family
counseling. Thus, even for black parents who are diagnosed as
having a mental illness, having limited access to treatment and
support services implies a lower likelihood of treatment seeking
behaviors, and thus mental illness among blacks may be seen as
less serious among both mental health and child welfare offi-
cials. Finally, it is noteworthy that when the full model is par-
titioned by race, the predictive efficacy for out-of-home place-
ment in the black model is 14 percent, but only 9 percent in the
white model. This finding illustrates that regardless of any ra-
cial bias which may exist within the child welfare system, it
appears different placement processes exist for black and white
children when deciding whether to remove a maltreated child
from the home.
Study Limitations
The sample used here is regional in nature, being limited to a
large urban county in the Northwest United States. Although
moderate in size, being comprised of 877 treatment cases, it is
hard to generalize these findings beyond the Northwest United
States to a national level without compromising their validity,
reliability and statistical significance. Another possible issue
with these data, as with any official data source, is the accuracy
of the files. Many courthouse officials and social workers who
document these child maltreatment cases may accidentally or
intentionally misreport the incidents of the child maltreatment
cases they are responsible for. Also problematic is the fact that
only those cases which became dependents of the court are
included in this study. Therefore, it is impossible to know if any
racial disparity in out-of-home placements can be accounted for,
either completely or partially, by a similarly large racial dispar-
ity in reporting.
Conclusion
This study aimed to answer two questions: First, does a mal-
treated child’s race affect their likelihood of receiving an out-
of-home placement? Second, do different factors matter for
black and white children within the decision making process?
In reference to the first question, two competing explanations
have been posited in previous literature. The first, black child-
ren experience more serious forms of maltreatment and have
fewer resources to remedy the maltreatment situation through
informal means than do white children. The second, there is an
underlying bias within the child welfare system, where dis-
criminatory beliefs about the perceived threat and dangerous-
ness of certain groups and their abilities to care for their chil-
dren may contribute to black children being disproportionately
removed from their homes. While this study finds evidence
supportive of the first explanation, the results are also consis-
tent with the possibility of an underlying racial bias within the
sys te m. The inclusion of a variety of maltreatment, demo-
graphic, familial and other risk factors accounted for approxi-
mately 28% of the racial disparity in out-of-home placements.
But, there remains a large amount disparity left to be explained.
While there always exists the possibility that some variables
with predictive power are not available in the data and are thus
left unexamined, that black children continue to experience
odds 2.404 times of white children in the full regression exam-
ined here lends support to the possibility of an underlying racial
bias within child welfare.
In reference to the second question, findings illustrate that
while few factors seem to matter differently for black and white
children to a statistically significant degree, there are two dif-
ferences that stand out. First is the differential effect of parental
mental illness, which is a significant predictor of white but not
black children being removed from the home. Second is the
differential effect of emotional abuse, in comparison to neglect,
which is a significant predictor of home removal for black child-
ren, but not white children. These findings seem to support the
conclusions made in previous research that black families may
be held to different standards, as compared to white families.
However, some caution may be necessary in the interpretation
of these findings. This study was unable to account for factors
such as parental awareness of the maltreatment of their children
or their amenability toward resolving the situation, which may
be key in determining aspects of parent’s mental and emotional
state.
Perhaps more important are those significant factors within
the partitioned models than those significantly different across
the models by race. For black children, the only predictive fac-
tors for the variation in out-of-home placement are the per-
centage of the population within the subject’s census tract re-
ceiving AFDC payments and emotional abuse as the maltreat-
ment type. For white children, parental mental illness, prior
childhood behavioral problems, sibling victimization and sex-
ual abuse matter to a statistically significant degree in deciding
whether out-of-home placement is warranted. Furthermore,
while the predictive efficacy for the model including only black
children was 14 percent, this same model had a predictive effi-
cacy of only 9 percent for white children. It appears for at least
black children, then, the child welfare system operates less on
the premise of removing a child from a home on the basis of
maltreatment, or more specifically in reference to specific
forms of abuse, than it does on the basis of what it sees as
long-term neglect or an inability to care. Among white children,
on the other hand, deleterious parental characteristics and abuse
seem to play a bigger role in deciding whether or not a mal-
treated child will be removed from the home. This is counter to
the present study’s hypothesis that familial risk factors would
Open Access.
326
A. M. KAUFMAN
exert a stronger effect among blacks than whites. A possible
explanation for these contradicting findings is that black fami-
lies are often assumed to possess these deleterious characteris-
tics, and thus, when such characteristics are actually confirmed
in a maltreatment case, they add little weight to the placement
decision. However, these characteristics, if viewed as less com-
mon among whites, would illustrate a more problematic living
situation, making out-of-home placement a more viable solu-
tion. Even these characteristics, though, seem to predict only a
small portion of variability in placements among white child-
ren, leaving it somewhat unclear what factors are important for
whites, or leading to the possible explanation that the decision
to remove white children from the home is made on a much
more individual-level basis than is the case for black families.
In light of these findings, CPS should initiate programs for
social workers to be sensitive to the specific issues that black
families face, such as single-parenthood and lower socioeco-
nomic status. It may also be in the best interest of these children
for CPS to lobby for further funding to provide such things as
mental counseling, drug abuse services, job placement, and
child care. Black families would then not be separated because
parents do not have the financial resources to provide for their
children according to the specific standards set by CPS or other
government officials. As much prior literature on out-of-home
placement and childhood outcomes has shown, removing a
child from the home can have long-term effects in many areas
of their lives, such as the ability to form attachments with oth-
ers, socioeconomic attainment, and risks for delinquency and
criminality (Currie & Widom, 2010; DeGue & Widom, 2009;
Ryan et al., 2008). Thus, whenever possible and in the best
interest of the child, the goal for CPS should be to seek what-
ever avenues necessary to keep a child with their family.
Future research analyzing the racial disparity in out-of-home
placements should examine to what extent these results can be
generalized to the entire population of maltreated children.
More qualitative and survey-oriented research may also be
helpful in understanding exactly what processes are at play to
make such factors as parental mental illness and parental sub-
stance abuse operate differently among black and white mal-
treated youth. This study was also not able to account for the
possible existence of reporting differentials, which may result
in a possible selection effect when analyzing just those cases
which make it to the official processing stage. Finally, a num-
ber of key variables which may further explain the racial-dis-
parity in out-of-home placement, but which were not available
in the present data, are family structure, a better measure of
maltreatment severity beyond the separation of maltreatment
types, and a more precise individual-level measure of socio-
economic status. Thus, it is imperative for researchers in the
child welfare arena to identify what additional factors contrib-
ute to the placement decision among maltreated children, and if
such characteristics can lead to a better understanding of the
seemingly differential processes for blacks and whites within
the child welfare system.
REFERENCES
Albonetti, C. A. (1991). An integration of theories to explain judicial
discretion. Social Problems, 38, 247-266.
http://dx.doi.org/10.2307/800532
Barth, R. P. (2009). Preventing child abuse and neglect with parent
training: Evidence and opportunities. Fut ur e o f Ch i l dr e n, 1 9, 95-118.
http://dx.doi.org/10.1353/foc.0.0031
Barth, R. P., Wildfire, J., & Green, R. L. (2006). Placement into foster
care and the interplay of urbanicity, child behavior problems and
poverty. America n Journal of Orthopsychiatry, 7, 358-366.
http://dx.doi.org/10.1037/0002-9432.76.3.358
Billingsley, A., & Giovannoni, J. (1972). Children of the storm: Black
children and American child welfare. New York: Harcourt, Brace
and Jovanovich.
Bridges, G. S., & Steen, S. (1998). Racial disparities in official assess-
ments of juvenile offenders: Attributional stereotypes as mediating
mechanisms. American Sociological Review, 63, 554-570.
http://dx.doi.org/10.2307/2657267
Brown, K. (2008). African American children in foster care: HHS and
congressional actions could help reduce proportion in care. Testi-
mony before the Subcommittee on Income Security and Family
Support, Committee on Ways and Means, House of Representatives.
United States Government Accountability Office.
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 & Ado-
lescent Psychiatry, 43, 960-970.
http://dx.doi.org/10.1097/01.chi.0000127590.95585.65
Courtney, M. E., McMurtry, S. L., & Zinn, A. (2004). Housing prob-
lems experienced by recipients of child welfare services. Child Wel-
fare, 83, 393-422.
Currie, J., & Widom, C. S. (2010). Long-term consequences of child
abuse and neglect on adult economic well being. Child Maltreatment,
15, 111-120. http://dx.doi.org/10.1177/1077559509355316
DeGue, S., & Widom, C. S. (2009). Does out-of-home placement me-
diate the relationship between child maltreatment and adult criminal-
ity? Child Maltreatment, 14, 344-355.
http://dx.doi.org/10.1177/1077559509332264
Drake, B., Lee, S. M., & Jonson-Reid, M. (2009). Race and child mal-
treatment reporting: Are blacks overrepresented? Children & Youth
Services Review, 31, 309-316.
http://dx.doi.org/10.1016/j.childyouth.2008.08.004
Hampton, R. L., & Newberger, E. H. (1985). Child abuse incidence and
reporting by hospitals: Significance of severity, class and race. Amer-
ican Journal of Public H ealth, 75, 56-60.
http://dx.doi.org/10.2105/AJPH.75.1.56
Hill, G. D., Harris, A. R., & Miller, J. L. (1985). The etiology of bias:
Social heuristics and rational decision making in deviance processing.
Journal of Research in Crime and Delinquency, 22, 135-162.
http://dx.doi.org/10.1177/0022427885022002003
Hill, R. B. (2004). Institutional racism in child welfare. Race and Soci-
ety, 7, 17-33. http://dx.doi.org/10.1016/j.racsoc.2004.11.004
Kirk, R. S., & Griffith, D. P. (2008). Impact of intensive family presser-
vation services on disproportionality of out-of-home placement of
children of color in one state’s child welfare system. Child Welfare
League of America, 87, 87-105.
Knott, T., & Donovan, K. (2010). Disproportionate representation of
African-American children in foster care: Secondary analysis of the
National Child Abuse and Neglect Data System, 2005. Children &
Youth Services Review, 32, 679-684.
http://dx.doi.org/10.1016/j.childyouth.2010.01.003
Lane, W. G., Rubin, D. M., Monteith, R., & Christian, C. W. (2002).
Racial differences in the evaluation of pediatric fractures for physical
abuse. Journal of the American Medical Association, 288, 1603-
1609. http://dx.doi.org/10.1001/jama.288.13.1603
Lindsey, D. (1994). Factors affecting the foster care placement decision:
An analysis of national survey data. American Journal of Orthopsy-
chiatry, 6, 272-281.
Martin, J. A., Hamilton, B. E., Sutton, P. D., Ventura, S. J., Menacker,
F., Kirmeyer, S., & Mathews, M. S. (2009). Births: Final data for
2006. Hyattsville, MD: National Center for Health Statistics.
Miller, O. A., & Gaston, R. J. (2003). A model of culture-centered child
welfare practice. Child Welfare, 82, 235-249.
National Poverty Center, The University of Michigan (2007). Poverty
in the United States. http://npc.umich.edu/poverty/
Open Access 327
A. M. KAUFMAN
Open Access.
328
Needell, B., Shaw, T., Magruder, J., & Putnam-Hornstein, E. (2007).
Racial disproportionality and disparity in child welfare: The dispar-
ity index. The National Child Welfare Data and Technology Confer-
ence, Washington, 18-19 July 2007.
Osterling, K. L., D’Andrade, A., & Austin, M. J. (2008). Understanding
and addressing racial/ethnic disproportionality in the front end of the
child welfare system. Journal of Evidence-Based Social Work, 5,
9-30. http://dx.doi.org/10.1300/J394v05n01_02
Phillips, S. D., Burns, B. J., Wagner, H. R., & Barth, R. P. (2004).
Parental arrest and children involved with child welfare service agen-
cies. American Journal of Orthopsychiatry, 74, 174-186.
http://dx.doi.org/10.1037/0002-9432.74.2.174
Pinderhughes, E. E. (1991). The delivery of child welfare services to
African American clients. American Journal of Orthopsychiatry, 61,
599-605. http://dx.doi.org/10.1037/h0079293
Polansky, N. A., Chalmers, M. A., Buttenweiser, E., & Williams, D. P.
(1981). Damaged parents: An anatomy of child neglect. Chicago:
The University of Chicago Press.
Rawal, P., Romansky, J., Jenuwine, M., & Lyons, J. S. (2004). Racial
differences in the mental health needs and service utilization of youth
in the juvenile justice system. The Journal of Behavioral Health Ser-
vices & Research, 31, 242-254.
Rivaux, S. L., James, J., Wittenstrom, K., Baumann, D., Sheets, J.,
Henry, J., & Jeffries, V. (2008). The intersection of race, poverty and
risk: Understanding the decision to provide services to clients and to
remove children. Child Welfare League of America, 87, 151-168.
Ryan, J. P., Testa, M. F., & Zhai, F. (2008). African American males in
foster care and the risk of delinquency: The value of social bonds and
permanence. Child Welfare League of America, 87, 115-140.
Sampson, R. J. (1986). Effects of socioeconomic context on official re-
action to juvenile delinquency. American Sociological Review, 51,
876-885. http://dx.doi.org/10.2307/2095373
Sampson, R. J., & Laub, J. (1993). Structural variations in juvenile
court processing: Inequality, the underclass, and social control. Law
and Society Review, 27, 285-311.
http://dx.doi.org/10.2307/3053938
Schuck, A. M. (2005). Explaining the black-white disparity in mal-
treatment: Poverty, female-headed families, and urbanization. Jour-
nal of Marriage and Family, 67, 543-551.
http://dx.doi.org/10.1111/j.1741-3737.2005.00152.x
Sedlack, A. J., & Broadhurt, D. D. (1996). Third National Incidence
Study of Child Abuse and Neglect. Washington, DC: US Department
of Health and Human Services, National Center of Child Abuse and
Neglect.
Tonry, M. H. (1995). Malign neglect: Race, crime, and punishment in
America. New York: Oxford University Press.
US Department of Health and Human Services, Administration on
Children, Youth and Families (2009). Child Maltreatment 2007.
Washington, DC: US Government Printing Office.
Widom, C. S., & Maxfield, M. G. (2001). An update on the “cycle of
violence”. Washington, DC: US Department of Justice Office of Jus-
tice Programs, National Institute of Justice.
Wulczyn, F. (2009). Epidemiological perspectives on maltreatment pre-
vention. The Future o f Children, 19, 39-66.
http://dx.doi.org/10.1353/foc.0.0029