See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/283110923

Demographics, policy, and foster care rates; A

Predictive Analytics Approach
ARTICLE in CHILDREN AND YOUTH SERVICES REVIEW NOVEMBER 2015
Impact Factor: 1.11 DOI: 10.1016/j.childyouth.2015.09.009

READS

14

2 AUTHORS, INCLUDING:
Jesse Rio Russell
National Council on Crime and Delinquency
18 PUBLICATIONS 18 CITATIONS
SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.

Available from: Jesse Rio Russell

Retrieved on: 26 February 2016

Children and Youth Services Review 58 (2015) 118126

Contents lists available at ScienceDirect

Children and Youth Services Review

journal homepage: www.elsevier.com/locate/childyouth

Demographics, policy, and foster care rates; A Predictive

Analytics Approach
Jesse Russell a,, Stephanie Macgill b
a
b

National Council on Crime and Delinquency, Madison, WI, United States

Best Friends Animal Society, Kanab, UT, United States

a r t i c l e

i n f o

Article history:
Received 18 November 2014
Received in revised form 13 September 2015
Accepted 14 September 2015
Available online 16 September 2015
Keywords:
Child maltreatment
Foster care
Foster care entry
Time in care
Child welfare policy
Demographics
Risk factors
Maltreatment risk factors
Classication and regression trees
CARTs

a b s t r a c t
Individual, family, and community-level factors have been suggested as explanations of foster care entry rates
and average lengths of time that children remain in foster care. They do not, however, provide a sufcient explanation of the substantial geographical variation in entry rates and average lengths of stay across the United States.
State-level child welfare policies and state-level socioeconomic variables may help explain these trends, but no
empirical analysis to date has identied how policies and socioeconomic facts might interact in ways that can
help account for the wide geographic differences.
Traditional statistical methods and much prior research have been unable to identify the combinatorial and nonlinear interaction effects of the many suggested factors. A data set of 104 state-level variables was constructed to
help answer the question of what accounts for geographic differences in foster care entry rates and average
lengths of stay in foster care. A predictive analytics approach (classication and regression trees) was used to
sort through all the potential explanatory variables, their interactions, and combinations. The results show that
state cultural orientations and socioeconomic facts together best explain foster care entry rates. In contrast,
child welfare policy and practice differences together best explain average lengths of stay in foster care. Interventions aimed at goals relating to who goes into foster care and how many children go into foster care might be
most effective if they focus on culture and socioeconomic facts. Interventions aimed to change lengths of time
in care, on the other hand, might be most effective if targeted at state child welfare policies and practices.
2015 Published by Elsevier Ltd.

1. Introduction
There is a gap in the research literature regarding how child welfare
policy along with individual, family, and community-level risk factors
might all work together to help explain geographic variations in foster
care entry rates and length of time in care across the United States.
This paper provides a better understanding of the differences in foster
care experiences across the United States by exploring what factors in
combination most efciently predict foster care entry rates and what
factors in combination are the most efcient predictors of the amount
of time that children spend in foster care.
The extant literature on foster care experiences might be best understood in terms of competing and accumulating risks: individual, family,
community, and cultural. So many variables have some potential bearing that traditional methodologies cannot account for them all. This
paper presents results from a nonlinear, nonparametric predictive analytics approach that allows for multiple combinatorial, nonlinear interactions. This type of predictive analytics modeling is the best way to
sort through all factors in order to identify those that empirically matter
the most.
Corresponding author at: 426 S. Yellowstone Dr., Madison, WI 53719.

http://dx.doi.org/10.1016/j.childyouth.2015.09.009
0190-7409/ 2015 Published by Elsevier Ltd.

There is considerable geographic variance in foster care entry rates

and in how long children remain in care across states. For example, in
2010, the foster care entry rate in Indiana was 5.7 per 1000 children
and in North Carolina it was 2.1. Put another way, children in Indiana
were nearly three times as likely as children in North Carolina to be
placed into foster care in 2010. The use of foster care as a state intervention was clearly different between those two states.
Foster care entry rates have a tendency to vary. For example, the foster care entry rate increased from 4.7 per 1000 children in 1980 to 7.7
per 1000 children in 2000 (Wertheimer, 2002). The foster care entry
rate trend later reversed, falling to 3.6 per 1000 children in 2010 (U.S.
Department of Health and Human Services, 2011).
Similarly, there is considerable variation in the average time that
children spend in foster care across the United States. In 2010, children
remained in care an average of 630 days (20 months) once removed
from the home, not counting the length of any previous foster care episodes. This average is consistent with length-of-stay trends since the
1990s (Barbell & Freundlich, 2001); however, how long children stay
in foster care varies signicantly from state to state. In Connecticut, children who exited foster care in 2010 had been in care an average of
810 days (almost 27 months), while children in New Jersey who exited
in 2010 had been in care an average of 614 days (20 months). Children

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

in Illinois exiting care in 2010 had been in care an average of 1336 days
(44 months; U.S. Department of Health and Human Services, 2010).
These inconsistencies across states in average lengths of stay in foster care are troubling because there is no clear standard for the level of
appropriate state response to child maltreatment. As one analyst has
suggested, while more aggressive use of foster care might increase
child safety, removal from parents can be traumatic to children as well
(Doyle, 2007). The wide differences across states in average lengths of
stay indicate that there is clear way to balance removing a child from
an unsafe environment and the trauma experienced by children from
lingering in foster care. Not having a complete understanding of state
variations in how this balance is made can hinder interventions aimed
at improving outcomes related to foster care entry and lengths of stay.
2. Potential explanations and variables
Potential explanations for these variations across states in foster care
have been suggested at the individual, the family, and the community
levels.
On the individual level, children with special needs that increase
caregiver burden and children younger than four years of age face an increased risk of abuse or neglect (Centers for Disease Control and
Prevention, 2012). A child's race may also be correlated with a greater
likelihood of foster care placement. In almost every state, African
American children are overrepresented in foster care when compared
with the general population (Summers, Wood, & Russell, 2013). A number of parental characteristics, such as age and level of education, also
place children at higher risk. Further, parental lack of understanding of
children's needs, child development and parenting skills, a parent's
own history of maltreatment, and parental substance abuse and mental
illness are all risk factors (Centers for Disease Control and Prevention,
2012; Hines, Lemon, Wyatt, & Merdinger, 2004).
On the family level, many factors have been found to be potentially
associated with child maltreatment risk. The factors, of course, vary by
family, but may include parental incarceration; low socioeconomic status; having more than four children; social isolation; family disorganization, dissolution, and violence; parenting stress; poor parentchild
relationships; and negative interpersonal interactions (Centers for
Disease Control and Prevention, 2012; Sedlak et al., 2010; World
Health Organization, 2002). Parents who were themselves maltreated
as children are also at greater risk of maltreating their own children
(Dixon, Browne, & Hamilton-Giachritsis, 2005).
On the community level, there are a number of factors that relate to
increased child maltreatment risk (Garbarino & Crouter, 1978). Lery
(2009) found that elements of neighborhood social structurein particular, residential instability, impoverishment, and child care burdenwere
related to the risk of entry into foster care. Community violence, acute
neighborhood disadvantage (e.g., high poverty and high unemployment
rates), poor social capital, and high population density are also related
to child maltreatment risk (World Health Organization, 2002).
Beyond community structure, culture or common community
values might relate to child maltreatment risk. For example, the prevalence of collectivist values (interdependence, social support, and a we
mindset) in a community versus prevalence of individualist values (independence, self-reliance, and an I mindset) could shape responses to
maltreatment incidents (Vandello & Cohen, 1999). A critical aspect of
individualism and collectivism for the purposes of this study is the different emphasis on relationships in each construct. Individualists view
relationships and groups as impermanent and non-intensive and
through a costbenet lens. Collectivists, on the other hand, see relationships and group memberships as permanent, stable, relatively impermeable, and important (Oyserman, Coon, & Kemmelmeier, 2002).
The role of culture has been examined within the child maltreatment
context, yet further understanding of its role is necessary (see, for example, Elliott & Urquiza, 2006; Stoltenborgh, Bakermans-Kranenburg, van
Ijzendoorn, & Alink, 2013).

119

The individualism-collectivism index used here from Vandello and

Cohen (1999) is based on an extensive research history of cultural
social-psychology constructs. These construct have be found to help explain patterns in behaviors, cognition, attitudes, goals, values and family
structures (see Triandis, 1996). Vandello and Cohen summarize the
construct as collectivism can be dened as a social pattern of closely
linked individuals who dene themselves interdependent members of
a collective (e.g., family and coworkers), whereas individualism as a cultural pattern stresses individual autonomy and individualism of the
self (p. 279). The index used is based on eight items: (1) percentage
of people living alone (reverse scored); (2) percentage of elderly people
(aged 65 and over) living alone (reverse scored); (3) percentage of
households with grandchildren living in them; (4) divorce to marriage
ratio (reverse scored); (5) percentage of people with no religious afliation (reverse scored); (6) average percentage number of people voting
Libertarian over the last four presidential elections (reverse scored);
(7) ratio of people carpooling to work to people driving alone; and
(8) percentage of self-employed workers (reverse scored).
An alternative explanation for the differences in foster care experiences may be variations in child welfare policy across states, which
can have an important effect on children's risk of maltreatment. Child
welfare policies vary from state to state, though all are built on the structure provided through federal law. States differ primarily in how they
have implemented laws and sometimes in laws themselves (in denitions of abuse and neglect, for example).
Federal laws (e.g., the Child Abuse Prevention and Treatment Act,
the Adoption Assistance and Child Welfare Act, the Adoption and Safe
Families Act, the Fostering Connections to Success and Increasing Adoptions Act, the Indian Child Welfare Act, and the Multiethnic Placement
Act) provide an important statutory framework regarding how states
ensure children's safety, permanency, and well-being. However, states
differ in the nature of implementation of these laws. For example, states
vary in terms of mandatory reporter laws and statutory denitions of
child abuse and neglect, and not all have extended foster care to youth
who are past the age of 18.
3. Current study
The current study addresses the gap in literature regarding how
child welfare policy and risk factors work together to explain variations
from state to state in foster care entry rates and length of time in care.
This study seeks to better understand the differences in children's
child welfare system experiences across the United States by exploring
the following research questions.
Research Question 1: What factors most efciently predict foster
care entry rates across states?
Research Question 2: What factors most efciently predict the
amount of time children spend in foster care across states?
The extant literature on foster care experiences might be best understood in terms of competing and accumulating risks: individual, family,
community, and cultural. So many variables have some potential bearing that traditional methodologies cannot account for them all. A nonlinear, nonparametric model that allows for multiple combinatorial,
nonlinear interactions is the best way at this stage of the research program to sort through all factors in order to assess those that matter
the most.
4. Method
4.1. Participant characteristics
The individual states in the United States were selected as the unit of
analysis. The most recent data up to 2010 were used when possible. The
District of Columbia was excluded from the data set because it was an

120

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

outlier along several variables, such as average income and average

child age, and because it was missing data for a number of variables.
4.2. Data set
State-level data on child welfare policies, outcomes, and demographics were collected from 10 sources: the Child Welfare Information
Gateway State Statutes Search (U.S. Department of Health and Human
Services, nda); the Child Welfare Enacted Legislation Database
(National Conference of State Legislatures, 2012); the KIDS COUNT
Data Center (Annie E. Casey Foundation, nd); the Reports and Results
of the Child and Family Service Reviews (U.S. Department of Health
and Human Services, ndb); the Fostering Connections Legislation by
State Database (Fostering Connections Resource Center, nd); the State
Child Welfare Policy Database (Casey Family Programs, 2010); the Welfare Rules Database (The Urban Institute, 2011); Census 2010 (U.S.
Census Bureau, nd); and the Adoption and Foster Care Analysis and
Reporting System (AFCARS) (U.S. Department of Health and Human
Services, 2010). Social values for each state were taken from Vandello
and Cohen (1999).
4.3. Measures
4.3.1. Foster care entry rate
The foster care entry rate was calculated using 2010 AFCARS and
2010 census data. AFCARS is a federal reporting system maintained by
DHHS that collects case-level information on all children in foster care.
In the AFCARS database, each child is counted only once and the information included on the child in the database is from the child's most recent foster care episode (U.S. Department of Health and Human
Services, 2012; U.S. Department of Health and Human Services, 2010).
For each state, the total number of foster care entries was divided by
the total child population and multiplied by 1000 to create a foster
care entry rate per 1000 children in each state.

predictor variables might interact. CARTs for the current study were created using the open source R statistics and programming platform. The
particular package used was the RPART routine, which is based on the
specications of Therneau and Atkinson (1997a, 1997b). While this statistical approach is not often found in social science scholarship, it is an
ideal method for the current study. The advantage of the approach is
that it reveals the contingent relationships that help to determine the
dependent variables. In addition, the approach can tell us which variables are the most efcient predictors (Russell, 2011).
CARTs are created with an algorithm that seeks to maximize homogeneity across values of the dependent variable within resulting tree
nodes. For each point on the scale of each variable, the algorithm tests
how well that particular point on that particular variable could separate
all the observations in the data set into two more homogenous groups.
Once every point on every variable is tested, the algorithm selects the
one that best separates all the observations into homogenous groups.
Then, for each of these two new groups, the algorithm proceeds again
to identify which point on which variable could separate the group
into two further more homogenous subgroups.
The algorithm seeks a balance between the most homogenous
groups possible and over-tting the model. This is achieved by measuring additional model complexity (such as the number of splits) as a
tradeoff with additional homogeneity with each potential split (measured with something like the Akaike information criterion). Balance
is also achieved by dividing the full data set into randomly drawn subsamples, developing multiple models, and cross-validating them with
each other. In the end, the model can be judged according to how well
it parsimoniously and accurately classies all the observations in the
data set.

4.3.2. Average length of time in care

Average length of time in foster care was derived from 2010 AFCARS
data by calculating the length of a child's stay in foster care from the
date of their latest removal to their exit date. The start of the current episode could be in any year, and the exit was in scal year 2010.
4.3.3. Independent variables
A total of 104 independent variables were entered into a database,
relating to legal denitions of abuse, neglect, and domestic violence;
statutes regarding mandatory reporting, kinship care, foster care licensing, and parent representation; degrees of implementation of differential response systems and Fostering Connections; Child and Family
Services Review (CFSR) ratings; poverty indicators; U.S. Census Bureau
demographics; AFCARS data; Temporary Assistance to Needy Families
(TANF) rules; or social values. Some variables of note are explained in
more detail below. See Appendix A for a detailed presentation of independent variables and their sources.
4.4. Analysis
Classication and regression trees (CARTs) are a form of predictive
analytics using a recursive partitioning algorithm. CARTs attempt to accurately and reliably predict an outcome of a particular case based on
the characteristics of that case (Kass, 1980; Breiman, Friedman,
Olshen, & Stone, 1984). The CART algorithm searches for patterns
(which may be nonlinear) and combinatorial relationships in a set of
data. CARTs are particularly well suited for uncovering hidden nonlinear
structures and interactions in complex data sets. Further, CARTs are an
appropriate method for exploratory analysis where the number of predictor variables is greater than the sample size (Strobl, Malley, & Tutz,
2009) or when there is not a theoretical basis for hypothesizing how

Fig. 1. Collectivism was the single most efcient predictor of foster care entry rates among
states.

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

Fig. 2. Child welfare expenditures was the single most efcient predictor of length of time
in care rates among states.

5. Results
The results are presented in the trees in Figs. 1 and 2. It is important
to note that while the recursive partitioning algorithm sequentially
makes computations on all the explanatory variables and calculates
every possible split for each variable, the nal results rely on splits
made on only a few variables. The variables and splits represented in
Figs. 1 and 2 are the most efcient in their ability to predict average foster care entry rates and average length of time in care.
The tree diagrams (shown in Figs. 1 and 2) can be read from the top
down. The full sample of cases is present at the top. The sample is then
split according to a sequence of criteria. At each split a single criterion is
given. The criterion is based on one independent variable, with higher
values going one way and lower values going another way. For example,
a split criterion might be based on a particular variable having values
greater or equal to 14.5. Cases for which the variable has a value of 15
or greater would be split to the left, and cases for which the variable
has a value of 14 or less would be split to the right. This process of splits
continues down the tree.
At the bottom of the tree are nal nodes, which are not split any further. The number below the node title is the average foster care entry
rate or average length of stay in care for the states in that node. The n
size indicates how many states are in that particular node along with
the percentage of observations in that node (out of 50).
5.1. Foster care entry rate
Model 1 shows how the CART algorithm separated all 50 observations through a series of binary splits to create groups (nodes) with

121

similar foster care entry rates. The CART algorithm identied collectivism as the most effective primary split (Fig. 1). Vandello and Cohen
(1999) developed a collectivism index that ranked each state's collectivism score; higher scores (on a scale of 1 to 100) indicate more collectivism. In the rst split, states with collectivism scores lower than 40 move
to the right to Node 1. The average foster care entry rate for the states in
Node 1 is 5.7 per 1000.
State diversity score was identied as the second split in Model 1.
Each state's race data (the number of individuals in a particular racial
or ethnic group), taken from the 2010 census, were entered into an online ShannonWiener Diversity Index tool (Chang Bioscience, 2011).
The ShannonWiener Diversity Index is a term used in biology studies
and measures the rarity and commonness of species in a biological community with the formula H = - [pi * ln(pi)] -pi (pi). The online tool
generated a score reecting the state's level of diversity; higher scores
reect greater diversity. States with diversity scores less than .87
move to the right to Node 2 and have an average foster care entry rate
of 4.2.
States with higher diversity scores were then split into a third group
according to the percentage of children living in crowded housing.
Taken from the KIDS COUNT Data Center, this variable is the percentage
of children under age 18 living in households that have more than 1 person per room. States with 12.5% or more of children living in crowded
housing move to the right to Node 3 and have an average foster care
entry rate of 3.7, while states with less than 12.5% move to the left to
Node 4 and have an average foster care entry rate of 2.4.
The predicted foster care entry rates were compared with the actual
foster care entry rates (Table 1). Doing so demonstrates whether a state
is consistent with the other states in the same node or is an aberration.
While there are several states in Nodes 1 and 2 whose actual foster care
entry rates differ from the group average by over two points, there are
no notable outliers. There is a smaller range of foster care entry rates
in Nodes 3 and 4.
To examine the predictive validity of the trees, we examined the
amount of foster care reentry for each node. The average percentage
of children who had been in foster care prior to 2010 was calculated
for each node in each model and compared with each node's average
foster care entry rates. In Model 1, the average percentage of reentries
for Node 1 was 26.3%. In other words, an average of 26.3% of the children
living in the states in Node 1 had been in foster care before their current
episode. In Node 2, the average percentage of reentries was 22.9%. In
Nodes 3 and 4, the percentage of reentries was 19.8% and 18.7%, respectively. Overall, the highest reentry averages matched the nodes with the
highest entry rates. A Pearson correlation coefcient was calculated to
assess the relationship between foster care entry rates and levels of reentry. There was a positive correlation between the two variables (r =
.41, p = b .01), indicating a moderately strong relationship.
5.2. Time in care
Model 2 presents how all 50 observations can be divided through a
series of binary splits into groups (nodes) with similar length of time
in care values. In Model 2, the CART algorithm identied a state's child
welfare expenditures as the most effective ways to split the full sample
into higher and lower time in care values. Child welfare expenditures include the administration and operation of maltreatment prevention services for children and families, family preservation services, child
protective services, in-home services, out-of-home placements, and
adoption services (Casey Family Programs, 2010). States with child welfare expenditures of $551 million or greater move to the right (see
Fig. 2).
States with lower child welfare expenditures were split according to
the minimum age at which youth are eligible for supervised independent living. Each state has its own standard for the appropriate age at
which youth can become eligible for supervised independent living programs. This variable represents only the age at which youth may begin

122

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

Table 1
Model 1 results compared to actual values for each state.

Table 2
Model 1 results compared to actual values for each state.

State

Model 1
node

Node average foster care

entry rate

Actual foster care

entry rate

State

Model 2
node

Node average length of

time in care (days)

Actual average length of

time in care (days)

Washington
Montana
Colorado
Kansas
North Dakota
Oregon
Iowa
South Dakota
Wyoming
Nebraska
New Hampshire
Utah
Maine
Idaho
Ohio
Wisconsin
Michigan
Pennsylvania
Missouri
Tennessee
Vermont
Minnesota
Kentucky
Arkansas
Indiana
Rhode Island
West Virginia
Texas
New York
Louisiana
Hawaii
New Mexico
Mississippi
California
Florida
Nevada
Oklahoma
Alaska
Arizona
Virginia
Illinois
Delaware
Maryland
North Carolina
Georgia
New Jersey
Alabama
Connecticut
South Carolina
Massachusetts

1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4

5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
5.7
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.2
3.7
3.7
3.7
3.7
3.7
3.7
3.7
3.7
3.7
3.7
3.7
3.7
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4

4
4.3
4.7
4.9
5.5
5.6
6.5
7.2
7.3
7.4
1.8
2.6
2.8
3.2
3.4
3.4
3.5
3.6
4.2
4.2
4.3
4.4
5.4
5.6
5.7
6.5
7.6
2.4
2.9
3
3.4
3.4
3.4
3.6
3.6
4.3
4.7
4.8
4.8
1.5
1.7
2
2.1
2.1
2.2
2.3
2.7
3
3.1
3.9

Indiana
Ohio
Florida
New Jersey
Pennsylvania
Massachusetts
Washington
California
Texas
Virginia
Connecticut
Michigan
New York
Maryland
Illinois
Arkansas
Minnesota
Wyoming
Tennessee
New Mexico
Colorado
Utah
South Carolina
Louisiana
Iowa
West Virginia
Delaware
Missouri
Oregon
Oklahoma
South Dakota
North Dakota
Idaho
Kentucky
Hawaii
Mississippi
Rhode Island
Kansas
Alabama
Arizona
Georgia
Wisconsin
Nebraska
Nevada
Vermont
North Carolina
Alaska
Montana
New Hampshire
Maine

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4

761
761
761
761
761
761
761
761
761
761
761
761
761
761
761
521
521
521
521
521
521
521
521
521
521
521
521
521
521
521
525
525
525
525
525
525
525
525
525
686
686
686
686
686
686
686
686
686
686
686

486
560
562
614
619
626
659
769
773
799
810
838
873
1095
1336
354
377
377
413
421
468
473
497
534
551
559
585
683
745
784
359
472
479
483
523
528
563
643
680
484
587
591
611
647
685
694
771
776
804
896

participating in supervised independent living programs. States with eligibility ages 16 or younger move to the left to Node 2 and have an average length of stay in care of 521 days.
States that moved to the right in terms of supervised independent living eligibility were split according to their scores on CFSR Safety Outcome
1, that children are rst and foremost protected from abuse and neglect:
Of all children who were victims of substantiated or indicated child
abuse and/or neglect during the rst six months of the year, what percentage had another substantiated or indicated report within a sixmonth period? (U.S. Department of Health and Human Services, 2013).
Mandated by ASFA, the CFSRs are one of the ways in which the federal
government exercises oversight of the child welfare system. Evaluators
assess state performance along a broad range of systemic, family, and
child outcome measures to determine how well states are performing
(Allen & Bissell, 2004; U.S. Department of Health and Human Services,
2011). Safety Outcome 1 reects the recurrence of substantiated or

indicated maltreatment. The CFSR ranks each state's achievement of this

outcome as a percentage (0 to 100). States with scores greater than or
equal to 84% (e.g., higher conformity) on Safety Outcome 1 move to the
left to Node 3 and have average lengths of stay in care of 525 days. States
with less conformity move to the right to Node 4 and have average
lengths of stay of 686 days. As with Model 1, the node averages were compared with the actual state averages for Model 2 (Table 2). There are
states in each node whose average length of stay in care is greater or
less than the group average by as many as 200 days.
Foster care reentries were also examined for Model 2. In Node 1, the
average percentage of children who had reentered the system was
21.1%. In Node 2, reentry was 22.7%. Nodes 3 and 4 averaged 22.7%
and 21.2%, respectively. Reentry levels across nodes were, therefore, virtually equal. How long children remain in care does not appear to be related to the degree to which children reenter care. If reentries were
systematically related to how long children remain in care, we would

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

expect to see a more linear relationship between reentry volume and

average lengths of stay in care. Furthermore, the Pearson coefcient indicated no relationship between the two variables (r = .12, p = .41).
6. Discussion
The purpose of this study is to begin to better understand state variations in terms of foster care entry rates and how long children remain
in foster care. To that end, a data set with more than 100 child welfare
policy-related and demographic variables was constructed. The CART
algorithms revealed that different factors most efciently predict foster
care entry rates and average lengths of stay in care.
In terms of foster care entry rates, a state's score along the social
value spectrum was the primary split predicting foster care entry
rates. States that scored lower on the collectivism scale or higher on
the individualism scale have higher average foster care entry rates. In
other words, states with the most individualistic orientation remove
children with more frequency than collectivistic states. This can be understood when examining the differences between an individualistic
culture's emphasis on independence, self-reliance, and personal dispositions and a collectivistic culture's emphasis on interdependence, social
support, and situational attributes. Individualistic cultures are more apt
to consider the individual when something happens, while collectivistic
cultures are more apt to examine the role situational factors play in crises. For example, in an individualistic culture, a mother who has a drug
problem is personally responsible for those decisions and needs to have
her children removed so that she can help herself. In a collectivistic culture, a mother with a drug problem may use drugs because she has no
job, skills, or support system and removing the children may not help,
but appropriate services might.
Less-diverse states have the second highest average foster care entry
rates. The results indicate that states with smaller minority populations
remove children at higher rates than more diverse states. This nding is
in accord with Foster (2012), who found that states with larger African
American populations are less likely to take children into protective custody, and who points to the need for a better understanding of the role
of race in child welfare decision-making.
Finally, foster care entry rates in states that are both more collectivist
and more diverse are predicted by how many children live in crowded
housing. States with more children living in crowded housing have
higher entry rates than states with fewer children in crowded housing.
Understanding this variable as a measure of poverty helps explain this
nding and corroborates explanations of poverty driving foster care
entry.
Not only do foster care entry rates appear to be explained by a state's
collectivistic orientation, diversity, and level of poverty, but entry rates
are also moderately correlated with a state's level of reentry, suggesting
that these variables can also help bolster understanding as to why children reenter foster care. It could be that families in individualistic states
are not receiving enough services to prevent maltreatment recurrence
because the individualistic orientation places emphasis on individual
responsibility over the intensive social services often needed to address
the causes of child maltreatment (e.g., mental health issues and
substance abuse). Secondly, families in individualistic states may not

123

be receiving the same type of help in their services. For example, individualistic services may be more focused on changing the individual,
while services in collectivistic contexts may be more focused on changing an individual's surroundingsto say nothing of a state's scal context and ability to maintain a comprehensive menu of social services.
Average lengths of time in care in states with lower child welfare expenditures were then split according to the minimum age at which foster youth are eligible for supervised independent living programs,
which are designed to provide youth with the necessary skills for successful adulthood. States whose eligibility age is 16 or younger have
the lowest average foster care stays.
States that do not allow youth 16 or younger to enter supervised independent living programs were then split according to their performance along CFSR Safety Outcome 1: that children are rst and
foremost protected from abuse and neglect. States with better performance along this measure had the second-lowest average length of
stay in care: four days longer than states that allow 16-year-olds to
enter supervised independent living programs. On the other hand,
states with less conformity with this measure had an average length
of stay that was 161 days longer.
Conversely to Model 1, which explored foster care entry rates and
found demographic and cultural values to be most salient, the CART algorithm in Model 2 identied policy and child welfare system performance factors as the most efcient predictors. It would appear that
policy decisions do have consequences for how children experience
the child welfare system.
The question of who goes into foster carewhich children and how
many childrenmight be best understood in terms of factors relating to
the state broadly: collectivism, diversity, and housing trends. The who
question appears to be less about policy decisions or practice approaches. In contrast, questions about the foster care experience, particularly lengths of time in care, might be best understood from policy
decisions and practice orientations: expenditures, safety goals, and
age-dependent programs.
The implications of this difference might be rst that it is difcult for
policy and practice decisions to overcome underlying cultural and socioeconomic facts when pursuing goals related to who goes into foster
care or how many children go into foster care. Instead, these goals
might be best approached by targeting cultural understanding and socioeconomic progress. Second, the implication might be that lengths
of stay in foster care can be understood as an attribute of how child welfare systems operatetheir decisions, policies, and practice approaches.
Goals related to safely reducing the amount of time children remain in
foster care, for example, might be best approached through effective
practice.
7. Conclusion
This paper explored what factors most efciently predict foster care
entry rates and how long maltreated children tend to remain in foster
care. A variety of variables were shown to matter the most (community
values, demographics, poverty, expenditure levels, policy on older
youth, and performance measures), and foster care reentry volume
was correlated with foster care entry rates.

Appendix A. Variables and sources

Variable

Source

Foster care entry rate

2010 Adoption and Foster Care Analysis Reporting System

(continued on next page)

124

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

(continued)
Variable

Average length of stay in foster care

3
4
5

23
24

Total child welfare expenditures, FY 2010

Percent change in total expenditures from FY 2008
Denition of physical abuse includes acts/circumstances threatening child with harm
or creating substantial risk of harm
Denition of neglect includes failure to educate child
Abandonment included in abuse or neglect status denition
Medical neglect dened in statute
Denition of abuse/neglect includes emotional maltreatment
Neglect depth
State has specic denition of sex abuse
Specic denition of emotional abuse/mental injury
Parental substance abuse
Domestic violence dened in civil statutes
Civil statutes list specic acts constituting domestic violence
Domestic violence included in civil denition of child abuse
Domestic violence dened in criminal or penal code
Criminal statutes list specic acts constituting domestic violence
Number of exceptions
Status of differential response implementation
Number of distinct pathways/tracks for screened-in reports
State has an approved Title IV-E Guardianship Assistance Program plan under
Fostering Connections
State has an approved Title IV-E plan to extend care to older youth under Fostering Connections
Does state have current Fostering Connections legislation?

25

Number of Fostering Connections bills passed

26

Does the state have pending Fostering Connections legislation?

27

State uses narrow denition of kin as relatives (related by blood, marriage, or

adoption) being treated differently than other kin (godparents, etc.)
State uses broad denition of kin dened as relatives and other kin having the same
treatment in all engagement with the child welfare agency
Number of methods used by child welfare agency to locate kin
State has ongoing involvement with kinship caregivers in private kin arrangements
State allows kin to be used as placement options to divert from foster care
State has a pre-approval process to allow children to be placed into kinship homes
almost immediately after they are removed
Licensure options for kin caring for a child in state custody
State allows kin to pursue permanent legal guardianship
Guardianship payment compared to foster care payment
State policy gives preference to noncustodial parents over other relatives
State statutory descriptions of who is required to report suspected or known child
abuse or neglect
Other professionals who are specied in state statutes as mandatory reporters other
than more commonly noted professionals (e.g., social workers, teachers, and physicians)
Youth can remain in foster care after 18th birthday under Fostering Connections
Youth can remain in foster care after 19th birthday under Fostering Connections
Cutoff foster care eligibility age
Number of circumstances allowing youth to stay in care past 18 (e.g., youth is on
track to graduate high school or get GED and youth has disabilities or special needs)
State has requirements youth must comply with to remain in foster care after 18th
birthday (e.g., youth must be employed and youth must be enrolled in school)
Number of requirements youth must comply with to remain in foster care after 18th birthday
Number of circumstances allowing youth to stay in care past 19 (e.g., youth is on
track to graduate high school or get GED; youth has disabilities or special needs)
State has requirements youth must comply with to remain in foster care after 19th birthday (e.g., youth
must be employed; youth must be enrolled in school)
Number of requirements youth must comply with to remain in foster care after 19th birthday
Court retains legal jurisdiction over youth in care after 18th birthday
Court retains legal jurisdiction over youth in care after 19th birthday
Supervised independent living is a placement option for youth in foster care
Minimum age for supervised independent living eligibility
Number of eligibility requirements for supervised independent living (e.g., youth
must be enrolled in school and youth must be working)
Youth can reenter foster care after emancipation or discharge to independent living
Age at which foster youth become eligible for Chafee-funded services
State uses its own funds for independent living/transition services and supports
Number of state-funded services and supports (e.g., scholarships and housing subsidies)
Number of outcome items demonstrating high performance (rating of strength) on the Child and Family
Services Review (CFSR) second round
CFSR Safety Outcome 1: Children are rst and foremost protected from abuse and neglect

6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58

Source
(AFCARS),
2010 Census (http://factnder2.census.gov)
2010 Adoption and Foster Care Analysis Reporting System
(AFCARS)
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Calculated from variables 47
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Fostering Connections Resource Center, fosteringconnections.
org
Fostering Connections Resource Center, fosteringconnections.
org
Fostering Connections Resource Center, fosteringconnections.
org
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Casey Family Programs, State Child Welfare Policy Database
Children's Bureau, Administration for Children and Families
Children's Bureau, Administration for Children and Families

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

125

(continued)
Variable

Source
Children's Bureau, Administration for Children and Families
Children's Bureau, Administration for Children and Families
Children's Bureau, Administration for Children and Families

67

CFSR Safety Outcome 2: Children are safely maintained in their homes when possible and appropriate
CFSR Permanency Outcome 1: Children have permanency and stability in their living situations
CFSR Permanency Outcome 2: The continuity of family relationships and connections is preserved for
children
CFSR Well-Being Outcome 1: Families have enhanced capacity to provide for their
children's needs
CFSR Well-Being Outcome 2: Children receive appropriate services to meet their
educational needs
CFSR Well-Being Outcome 3: Children receive adequate services to meet their physical
and mental health needs
CFSR Number Systemic Factors in conformity (Information System, Quality
Assurance System, Service Array, etc.)
Number of circumstances in which reasonable efforts are not required
(e.g., instances when the parent's rights to another child have been terminated)
Number of child maltreatment victims in 2010

68
69
70
71
72
73
74

Percentage of children in poverty

Percentage of total population in poverty
Percentage of children in single-parent families
Percentage of children affected by foreclosure from 2007 to 2009
Percentage of parents without health insurance
Percentage of children living in crowded housing
Average monthly foster care payment

75

Percentage of male children in foster care

76

Percentage of female children in foster care

77

Average child age at start of scal year (October 1)

78
79
80
81
82

93
94
95
96
97

To become a foster parent, state requires training prior to licensure by law

State provides a specic course of training
Require other specialized training (e.g. CPR, rst aid, and re prevention)
Number of training hours required
Number of minimum standard safety requirements for foster homes (e.g., smoke
detector, carbon monoxide detector, and re extinguisher)
Home must have sufcient number of bedrooms so children of opposite sex do not share
Require a home inspection by the state health department or by a re marshal
State requires results of recent health exams
Federal background check required for applicant
Child abuse and neglect registry check required for applicant
Require check of child abuse and neglect registries in any other state where applicant has lived in past ve
years
Fingerprints required for prospective foster parents
Background checks required for all household members, regardless of age
State prioritizes relatives for out-of-home placements
Relative may be issued a temporary license but must be able to meet all requirements for full licensure
after temporary license expires
Relatives must meet all regulations for licensure before a child can be placed in their care
Child may be placed without formal licensing
Average income
Median income
Diversity score

98

Collectivism score

99

Parent representation in dependency cases is a state due process right (afrmed by a

court decision)

59
60
61
62
63
64
65
66

83
84
85
86
87
88
89
90
91
92

100 Parent representation in dependency cases statutorily required

Children's Bureau, Administration for Children and Families

Children's Bureau, Administration for Children and Families
Children's Bureau, Administration for Children and Families
Children's Bureau, Administration for Children and Families
Casey Family Programs, State Child Welfare Policy Database
Child Welfare Outcomes Report, Children's Bureau,
Administration for Children and Families
KIDS COUNT Data Center, datacenter.kidscount.org
KIDS COUNT Data Center, datacenter.kidscount.org
KIDS COUNT Data Center, datacenter.kidscount.org
KIDS COUNT Data Center, datacenter.kidscount.org
KIDS COUNT Data Center, datacenter.kidscount.org
KIDS COUNT Data Center, datacenter.kidscount.org
The Adoption and Foster Care Analysis and Reporting System
(AFCARS),
Children's Bureau, Administration for Children and Families
The Adoption and Foster Care Analysis and Reporting System
(AFCARS),
Children's Bureau, Administration for Children and Families
The Adoption and Foster Care Analysis and Reporting System
(AFCARS),
Children's Bureau, Administration for Children and Families
The Adoption and Foster Care Analysis and Reporting System
(AFCARS),
Children's Bureau, Administration for Children and Families
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
Child Welfare Information Gateway State Statutes Search
2010 American Community Survey, United States Census
2010 American Community Survey, United States Census
Calculated with data from American Fact Finder, United
States Census Bureau
Vandello, J. A., & Cohen, D. (1999). Patterns of individualism
and collectivism
across the United States. Journal of Personality and Social
Psychology, 77(2), 279292.
Sankaran, V. (nd). A national survey on a parent's right to
counsel in termination of
parental rights and dependency cases. Retrieved from:
http://www.law.umich.edu/
centersandprograms/ccl/specialprojects/Documents/National
%20Survey%
20on%20a%20Parent's%20Right%20to%20Counsel.pdf
Sankaran, V. (nd). A national survey on a parent's right to
counsel in termination of
parental rights and dependency cases. Retrieved from:
http://www.law.umich.edu/
centersandprograms/ccl/specialprojects/Documents/National
%20Survey%
(continued on next page)

126

J. Russell, S. Macgill / Children and Youth Services Review 58 (2015) 118126

(continued)
Variable
101 Parent representation is afforded in statute but is also discretionary

References
Allen, M., & Bissell, M. (2004). Safety and stability for foster children: The policy context.
The Future of Children, 14(1), 4873.
Annie E. Casey Foundation (nd). KIDS COUNT data center. Retrieved from http://www.
datacenter.kidscount.org.
Barbell, K., & Freundlich, M. (2001). Foster care today. Washington, DC: Casey Family Programs (pp. 17).
Breiman, L., Friedman, J. H., Olshen, R. A., & Stone, C. J. (1984). Classication and regression
trees. Belmont, CA: Wadsworth Publishing.
Casey Family Programs (2010). The state child welfare policy database. Retrieved from
http://www.childwelfarepolicy.org/maps.
Centers for Disease Control and Prevention (2012). Child maltreatment: Risk and protective
factors.
Retrieved
from
http://www.cdc.gov/violenceprevention/
childmaltreatment/riskprotectivefactors.html.
Chang Bioscience (2011). ShannonWiener diversity index/Shannon entropy calculator.
Retrieved from: http://www.changbioscience.com/genetics/shannon.htm.
Dixon, L., Browne, K., & Hamilton-Giachritsis, C. (2005). Risk factors of parents abused as
children: A mediational analysis of the intergenerational continuity of child maltreatment (part I). Journal of Child Psychology and Psychiatry, 46, 4757.
Doyle, J. J. (2007). Child protection and child outcomes: Measuring the effects of foster
care. American Economic Review, 97(5), 15831610.
Elliott, K., & Urquiza, A. (2006). Ethnicity, culture, and child maltreatment. Journal of Social
Issues, 62(4), 787809.
Foster, C. H. (2012). Race and child welfare policy: State-level variations in
disproportionality. Race and Social Problems, 4(2), 93101.
Fostering Connections Resource Center (nd). Legislation by state.
Garbarino, J., & Crouter, A. (1978). Dening the community context for parentchild relations: The correlates of child maltreatment. Child Development, 49, 604616.
Hines, A. M., Lemon, K., Wyatt, P., & Merdinger, J. (2004). Factors related to the disproportionate involvement of children of color in the child welfare system: A review and
emerging themes. Children and Youth Services Review, 26(6), 507527.
Kass, G. V. (1980). An exploratory technique for investigating large quantities of categorical data. Applied Statistics, 29(2), 119127.
Lery, B. (2009). Neighborhood structure and foster care entry risk: The role of spatial scale
in dening neighborhoods. Children and Youth Services Review, 31, 331337.
National Conference of State Legislatures (2012). Child welfare enacted legislation database. Washington, D.C.: Author.
Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological
Bulletin, 128(1), 3.
Russell, J. (2011). Hidden patterns in exchange rate regime choice. Empirical Economics,
40(2), 425449.
Sedlak, A. J., Mettenburg, J., Basena, M., Peta, I., McPherson, K., & Greene, A. (2010). Fourth
national incidence study of child abuse and neglect (NIS-4). Washington, DC: U.S. Department of Health and Human Services Retrieved from: http://www.law.harvard.
edu/programs/about/cap/cap-conferences/rd-conference/rd-conference-papers/
sedlaknis.pdf.
Stoltenborgh, M., Bakermans-Kranenburg, M. J., van Ijzendoorn, M. H., & Alink, L. R.
(2013). Culturalgeographical differences in the occurrence of child physical abuse?

Source
20on%20a%20Parent's%20Right%20to%20Counsel.pdf
Sankaran, V. (nd). A national survey on a parent's right to

A meta-analysis of global prevalence. International Journal of Psychology, 48(2),

8194.
Strobl, C., Malley, J., & Tutz, G. (2009). An introduction to recursive partitioning. Munich,
Germany: Department of Statistics, University of Munich.
Summers, A., Wood, S., & Russell, J. (2013). Disproportionality rates for children of color in
foster care (Technical Assistance Bulletin). Reno, NV: National Council of Juvenile and
Family Court Judges.
The Urban Institute (2011). The welfare rules database. Washington, DC: Author Retrieved
from http://anfdata.urban.org/wrd.
Therneau, T. M., & Atkinson, E. J. (1997a). An introduction to recursive partitioning using
the RPART routines (Technical Report 61, Section of Biostatistics). Retrieved from
Mayo Clinic website: http://www.mayo.edu/hsr/techrpt/61.pdf.
Therneau, T. M., & Atkinson, E. J. (1997b). An introduction to recursive partitioning using the
RPART routines. Technical Report 61. Rochester, MN: Mayo Clinic (p. 452).
Triandis, H. C. (1996). The psychological measurement of cultural syndromes. American
Psychologist, 51(4), 407.
U.S. Census Bureau (nd). Census 2010. American Fact Finder. Retrieved from http://
factnder2.census.gov/faces/nav/jsf/pages/index.xhtml.
U.S. Department of Health and Human Services (2010). Adoption and foster care analysis
and reporting system. Washington, D.C.: Administration for Children and Families,
Children's Bureau.
U.S. Department of Health and Human Services (2011). Federal child and family services
reviews: Aggregate report, round 2. Washington, D.C.: Administration for Children
and Families, Children's Bureau Retrieved from http://www.acf.hhs.gov/sites/
default/les/cb/fcfsr_report.pdf.
U.S. Department of Health and Human Services (2012). About AFCARS. Washington, D.C.:
Administration for Children and Families, Children's Bureau Retrieved from http://
www.acf.hhs.gov/programs/cb/resource/about-afcars.
U.S. Department of Health and Human Services (2013). Child welfare outcomes
20082011: Report to Congress. Washington, D.C.: Administration for Children and
Families, Children's Bureau Retrieved from http://www.acf.hhs.gov/programs/cb/
resource/cwo-08-11.
U.S. Department of Health and Human Services (nda). State statutes search. Washington,
D.C.: Administration for Children and Families, Children's Bureau, Child Welfare Information Gateway. Retrieved from https://www.childwelfare.gov/systemwide/
laws_policies/state/.
U.S. Department of Health and Human Services (ndb). Reports and results of the Child
and Family Service Reviews. Washington, D.C.: Administration for Children and Families, Children's Bureau. Retrieved from http://basis.caliber.com/cwig/ws/cwmd/docs/
cb_web/SearchForm.
Vandello, J. A., & Cohen, D. (1999). Patterns of individualism and collectivism across the
United States. Journal of Personality and Social Psychology, 77(2), 279292.
Wertheimer, R. F. (2002). Youth who age out of foster care: Troubled lives, troubling prospects (Child Trends Research Brief). Baltimore, MD: Annie E. Casey Foundation Retrieved from http://www.aecf.org/resources/youth-who-age-out-of-foster-caretroubled-lives-troubling-prospects/.
World Health Organization (2002). World report on violence and health. Geneva,
Switzerland: Author.