Educational and employment outcomes of adults Northwest Foster Care Alumni Study

Children and Youth Services Review
28 (2006) 1459 – 1481
Educational and employment outcomes of adults
formerly placed in foster care: Results from the
Northwest Foster Care Alumni Study
Peter J. Pecora a,b,⁎, Ronald C. Kessler c , Kirk O'Brien a ,
Catherine Roller White a , Jason Williams d , Eva Hiripi c ,
Diana English b,e , James White f , Mary Anne Herrick g
Casey Family Programs, United States
University of Washington, School of Social Work, United States
Harvard Medical School, United States
University of Alaska Anchorage, United States
Child Welfare League of America, United States
Portland State University, United States
Washington Education Foundation, United States
Received 3 February 2006; received in revised form 29 March 2006; accepted 7 April 2006
Available online 5 July 2006
This study evaluated the intermediate and long-term effects of family foster care on adult functioning
using a sample of 659 young adults from two public and one private child welfare agencies, case record
reviews, structured interviews, and a survey response rate of 76%. Foster care alumni completed high
school at a rate comparable to the general population, but a disproportionately high number of them
completed high school via a GED. Alumni completion rates for postsecondary education were low.
Consequently, many alumni were in fragile economic situations: one-third of the alumni had household
incomes at or below the poverty level, one-third had no health insurance, and more than one in five
experienced homelessness after leaving foster care. Two foster care experience areas were estimated to
significantly reduce the number of undesirable outcomes in the Education outcome domain: positive
placement history (e.g., high placement stability, few failed reunifications), and having broad independent
living preparation (as exemplified by having concrete resources upon leaving care). For the Employment
and Finances outcome domain, receiving broad independent living preparation (as exemplified by having
⁎ Corresponding author. Casey Family Programs, 1300 Dexter Ave. North, Floor 3, Seattle, WA 98109, United States.
Tel.: +1 206 270 4936; fax: +1 866 322 7863.
E-mail address: [email protected] (P.J. Pecora).
0190-7409/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
concrete resources upon leaving care) was estimated to significantly reduce the number of undesirable
© 2006 Elsevier Ltd. All rights reserved.
Keywords: Educational and employment outcomes; Foster care; Northwest Foster Care Alumni Study
1. Introduction
1.1. The scope of foster care in the United States
In any given year, about 800,000 children are served in foster care in family and non-family
settings, with a daily census of 518,000 (as reported at the end of September 2004; U.S. Department of Health and Human Services, 2005c). This number represents about 1% of children in
America.1 The numbers of children in foster care have risen substantially since 1980 and are only
now just leveling off (see Fig. 1). Most of these children have suffered maltreatment, although not
all are placed specifically because of abuse or neglect. Some are placed due to other parental
dysfunction, parent death, or because of child mental health service needs that the family can not
afford to pay for (U.S. Department of Health and Human Services, 2001).
Even though preventing the placement of children in foster care and minimizing their length of
stay is a child welfare priority, many children will spend a substantial amount of their childhood
living in foster care (U.S. Department of Health and Human Services, 2005a; Wulczyn, Barth,
Yuan, Harden, & Landsverk, 2005; Wulczyn & Goerge, 1992). Nearly half of the children who
are placed in foster care will remain there for a year or longer with an average length of stay of
2 years. More specifically, of those children still in foster care as of September 30, 2003, 61% had
been there for 12 months or longer. Of those leaving care in fiscal year 2003, 50% had been in
care 11 months or less, but 20% had been there for 3 years or more (U.S. Department of Health
and Human Services, 2005b).
Over 21,000 older youth emancipate to adulthood from a foster care setting every year (U.S.
Department of Health and Human Services, 2005b). To evaluate how youth have been impacted by
foster care, this article focuses on the adult education, employment and finance achievements of
young adults ages 20–33 years old who spent 12 months or more in family foster care as adolescents
during the 10 years prior to the implementation of the Foster Care Independence Act of 1999 (1988–
1998). The study examines foster care experiences that are associated with educational achievement
and a positive financial situation after leaving foster care.
1.2. Current goals, objectives, and key outcomes of foster care
Given the large number of children placed and the extended time that many spend in foster
care, systems have been instituted to provide for the well-being of these children. Current foster
care practice and programs are governed by an intricate–although not necessarily coherent–set of
This figure is derived using the following statistics: 800,000 children in foster care divided by the total number of
children under age 18 in the United States on July 1, 2003 (800,000/73,043,506 = 1.1%). For national foster care data, go
to and click on “Adoption and Foster Care Statistics.” See the “Trends in Foster
Care and Adoption” table. The total number of children under age 18 was downloaded February 21, 2006 from http:// (Annual Estimates of the Resident Population by Selected Age Groups for the United States and States:
July 1, 2003 and April 1, 2000).
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Fig. 1. Children in foster care: 1998–2004. Sources: Trends in Foster Care and Adoption Chart. AFCARS data, Children's
Bureau, Administration for Children and Families, HHS: based on data submitted September 2005 (http://www.acf.hhs.
gov/program/cb/), Voluntary Case Information System statistics compiled by the American Public Human Services
Association, and U.S. Census Bureau, Resident Population Estimates of the United States by Age and Sex: http://www. for 1988 through 1999 and:
html for 2000 through 2004.
policies and laws at the federal, state, and local levels (Curtis, Grady, & Kendall, 1999; Lindsey &
Schwartz, 2004; Pew Commission, 2004). According to professional and federal standards, child
welfare agencies need to achieve a small but critical set of outcomes in three broad outcome
domains: child safety, child permanence, and child and family well-being (U.S. Department of
Health and Human Services, 2003).
For children placed for a year or longer, more specific outcomes should be pursued related to
emotional, physical, and cognitive development. These outcomes may vary by age of the child
and length of stay, but could include reduction in the emotional trauma of child maltreatment;
healthy physical development through regular check-ups and adequate medical, dental, and vision
care; avoidance of teen pregnancy; life skills development and readiness for adulthood; high
school graduation and enrollment in post-secondary education or training programs; healthy
socialization and healthy adult relationships (American Academy of Pediatrics, 2000; Berrick,
Needell, Barth, & Johnson-Reid, 1998; Casey Family Programs, 2001, 2003a). Wulczyn et al.
(2005, p.180), however, have raised cautions about the concept of child well-being:
… the term well-being is really shorthand for health, education, and mental health, facets of
well-being that are for the most part programmatically beyond the scope of the child
welfare system. By using a vague term such as well-being when something much more
specific is intended, advocates for expanded services for children weaken their position.
The child welfare system cannot be accountable for safety, permanency, and education,
health, and mental health because the resources needed to influence the well-being outcomes are outside the boundaries of the system.
Instead, Wulczyn et al. suggest that the mandate for the child welfare system should be
expressed as achieving safety and permanency outcomes and ensuring the child welfare system
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
engages the education, health and mental health systems. This will help the child welfare system
rightfully take credit for working to create safe and stable families for children. It will also have a
second important benefit:
…separating well-being into its constituent parts makes clear that which is obvious to most
child welfare administrators: When well-being is couched in terms of health, mental
health, and education, the general public and policymakers alike quickly realize that the
child welfare system has a limited range of options without using the health care system,
schools, and the mental health system as collaborators. If children in the child welfare
system are not moving along their age-appropriate trajectory, these other systems will have
to shoulder their share of the responsibility in conjunction with, not in spite of, the child
welfare system (Wulczyn et al., 2005, p. 180).
1.3. How effective is foster care?
Despite the growing clarity about expected outcomes, little is known about the long-term
outcomes for youth in foster care. But accountability for the effects of foster care services is of
high concern, with pressure from private and public agencies, juvenile court judges, physicians,
national commissions, and various other stakeholder groups, as well as from class action lawsuits
(Pecora, Massinga, & Mauzerall, 1997; Pew Commission, 2004; United States General Accounting Office, 1999). Federal legislative action has focused on introducing key measures or
outcome indicators emphasizing such important areas as length of care, safety from child abuse or
neglect, placement history, and permanency (U.S. Department of Health and Human Services,
1999, 2001, 2003, 2005a).
Even with $22 billion of federal, state and local funding annually (Scarcella, Bess, Zielewski,
Warner, & Geen, 2004), there are reports that the foster care system is in crisis, that foster parent
retention is low, workers remain overloaded with high caseloads; and that poor salaries, inadequate
supervision and stressful working conditions result in high staff turnover and uneven job
performance (Pew Commission, 2004; U.S. General Accounting Office, 2004). While it is true that
there is some unevenness in agency performance, there are several major issues that need to be
addressed when examining the effectiveness of foster care. First, about 50% of the children placed
each year in foster care stay in care for less than 1 year—a very brief time for any one agency to
show dramatic improvements in various areas of child functioning. Second, it is difficult to
demonstrate that what happens to foster children is the result of being in foster care, and not merely
something which happens as a consequence of pre-determined characteristics of those entering the
foster care system:
Their often severe difficulties make it unsurprising that the rate of emotional, social,
behavioral and educational problems found in children in foster care is substantially higher
than in the general population. This is not to say that foster care should not do more to
improve these problems; merely that they pre-exist for most children and should not
necessarily be seen as evidence of a lack of effectiveness or the failures of the state to be a
good parent (Wilson, 2006, p. 7).
Notwithstanding Wilson's comments, there is growing evidence that certain aspects of foster
care negatively influence child outcomes such as maltreatment by foster parents and frequent
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
placement changes (Ryan & Testa, 2004). Conversely, there are data that youth placed in foster
care have similar (and in some areas) more positive adult outcomes, as compared to youth in the
general population from similar family backgrounds (Kessler & Magee, 1993; Buehler, Orme,
Post, & Patterson, 2000). Some studies indicate that children who receive foster care services
show positive improvements in physical health, emotional adjustment, school performance and
behavioral functioning compared to the years before being placed into care (Berrick et al., 1998;
Biehal & Wade, 1996; Coulling, 2000; Goerge, Wulczyn, & Fanshel, 1994). Other studies suggest
children in foster care compared to children from poor and ethnic minority families are at greater
risk for becoming involved in criminal activities, suffer from more frequent and debilitating
mental disorders (Courtney, Terao, & Bost, 2004; Minty, 1999), or when compared to the general
population, have higher homelessness and unemployment rates (Dworsky, 2005; Park, Metraux,
& Culhane, 2005).
Further complicating the ability to draw decisive conclusions about foster care effectiveness is
that few studies have been designed to allow for valid comparisons to children in the general
population. Additionally, use of brief follow-up periods (1 year after leaving care) and relatively
small sample sizes of research studies involving children in foster care (often less than 150 subjects)
prevent anyone from drawing consistent and generalizable conclusions (Barth, 1990; Cuddeback,
2004; Holdaway & Ray, 1992; Jacobson & Cockerum, 1976; Jones & Moses, 1984; Rice &
McFadden, 1988). Findings are simply inconclusive.
The ideal means to examine the effectiveness of foster care is to compare a random sample of
maltreated youth that were either removed from the home or not removed from the home (but
received in-home services). Because of the ethics involved in allowing children to remain in a
home where maltreatment has occurred, this design has been used rarely. Only one experimental
study compared the functioning of youth placed in foster care and youth served by alternative
means while they remained with their birth parents, but that evaluation was compromised severely
by sampling problems (Wald, Carlsmith, & Leiderman, 1988). Equally scarce are studies that
compared youth raised in foster care with a group of general population youth matched on certain
family or youth characteristics (Blome, 1997; Buehler, Orme, Post, & Patterson, 2000; Courtney et
al., 2004; Minty, 1999; Runyan & Gould, 1988), studies that compared the effects of various foster
parent payment and support patterns (Jaffee & Kline, 1970), and studies that have examined costs
or cost–benefits in foster care (Barth, 1997; Barth, Lee, Wildfire, & Guo, in press; Kerman, Barth
& Wildfire, 2004; Sedlak & Broadhurst, 1993; Young & Brandt, 1977).
Another methodological limitation has been the lack of precision and consistency in measuring
placement change, despite the importance of this variable for youth development. While research
has shown that some changes in living situation can be beneficial for youth, greater disruption in
placement has been associated with child behavior problems (Dore & Eisner, 1993; Newton,
Litrownik, & Landsverk, 2000; Ryan & Testa, 2004, 2005; Smith, Stormshak, Chamberlain, &
Whaley, 2001), and mental health problems (Rubin et al., 2004). For example, Newton et al.
(2000, p. 1371) found that volatile placement histories “contribute negatively to both internalizing
and externalizing behavior of foster children, and that children who experience numerous changes
may be at particularly high risk for these deleterious effects.”
In conclusion, a substantial number of youth enter foster care and are likely to remain there for
an extended period of time, and are cared for by a patchwork system of care, consisting of child
welfare agencies and local, state, and federal governments. Further, studies have demonstrated
mixed results and few rigorous studies have been conducted to date. Understanding the outcomes
of youth who have been placed in foster care has been encumbered by a lack of random assignment
studies and other methodological issues. To help address these needs, the Northwest Foster Care
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 1
Demographic characteristics of the Northwest alumni
Non-Hispanic black
Non-Hispanic other
Non-Hispanic white
Age at time of interview
20 to 22 years
23 to 25 years
26 to 33 years
Mean age at time of interview
Sample size a
Sample size
% (S.E.)
10.5 (0.9)
21.3 (1.2)
22.6 (1.2)
45.6 (1.5)
29.6 (1.4)
38.2 (1.4)
32.2 (1.4)
24.2 (0.1)
39.5 (0.7)
60.5 (0.7)
Age at time of interview and ethnicity report data drawn from alumni interviews.
Alumni Study was designed to evaluate how young adults who had been placed in foster care were
functioning. Three main research questions were examined for this article:
1. What were the educational achievements of the alumni?
2. What was the financial situation of these alumni?
3. What foster care experiences were associated with educational achievement and a positive
financial situation after leaving foster care?
1.4. The Northwest Foster Care Alumni Study
The primary purpose of the Northwest Foster Care Alumni Study was to evaluate the intermediate and long-term effects of family foster care on adult outcomes. The University of Washington, Harvard Medical School and the University of Michigan Survey Research Center collaborated
with three social service organizations to conduct the study:
1. Casey Family Programs, a private operating foundation, with participating offices in Seattle,
Tacoma and Yakima, Washington, and Portland, Oregon (abbreviated as Casey);
2. Oregon Department of Human Services, Children, Adults and Families, with participating
offices in Portland, Oregon; and
3. The State of Washington Department of Social and Health Services, Children's Administration,
Division of Children and Family Services, with participating offices in the areas of Seattle,
Tacoma and Yakima, Washington.
The investigation focused on adults who spent time in foster care in one of the three agencies from
1988 to 1998. There were a few differences in the child demographics and services received by the
foster care alumni from each agency. For example, Casey alumni were older, placed at a younger age,
and for a longer duration.2 To accommodate for these pre-existing differences, steps were taken to
Prior to entering Casey, nearly 9 in 10 (89.0%) of Casey alumni spent time in a state or a voluntary agency program.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
adjust for differences between the Casey and State samples, as well as to account for the youth who
were not interviewed. Propensity score matching was used to weight the data so that differences
between the agencies, and between those interviewed and those who were not, did not bias the results
(Braitman & Rosenbaum, 2002). In sum, weights were used to estimate data as if there had been no
pre-existing differences between Casey and the state samples, and as if the entire sample had been
interviewed. (This is why the full sample size of 659 is used for the demographics, risk factors and
placement history variables.) Variables that contributed to the statistical weights included age, sex,
and race. Combined, these techniques improved the ability to generalize to the population of youth
from which the sample was drawn.
2. Method
2.1. Participants
In addition to being placed by Casey or the state child welfare agency in Seattle, Tacoma and
Yakima, Washington or Portland, Oregon, participants had to (1) have been served between
January 1, 1988 and September 30, 1998; (2) have been placed for 12 months or more in family
foster care between the ages of 14 and 18 years old; (3) not have a major physical or developmental
disability (e.g., IQ scores of less than 70), and (4) have been placed due to reasons other than being
an unaccompanied refugee minor. A total of 659 alumni fit these criteria.
2.2. Demographics
Over half (54.4%) of the interviewed alumni were people of color (see Table 1). American
Indian or Alaska Native alumni (15.7%) and Asian or Native Hawaiian/Polynesian/Pacific
Islander alumni (5.8%) made up the largest proportion of the “Non-Hispanic other” ethnic group.
The average age of the alumni was 24.2 years old, and 60.5% were women.
2.3. Materials
2.3.1. Case records
Raters, with no knowledge of the hypotheses of the study, individually read and recorded the
information from case records. Only variables which had acceptable inter-rater reliability (as
determined by Cohen's Kappa or by inter-rater correlations) were retained. Case records provided
information on living arrangements before foster care, parent functioning, child maltreatment by
birth family, reasons for initial placement, mental and physical health problems, and placement
information. Further, the study also collected data about youth disabilities and diagnoses from case
records. Data focused on professionally-diagnosed medical, learning, and psychological issues
that could have surfaced at any point in childhood, including before placement and up to case
closure. Lastly, operational definitions of abuse and neglect variables as defined by the Barnett
Coding System for type and severity of child maltreatment, and modified for use by the LONGSCAN project (Multiple Method Classification System) were used to record child maltreatment
data (Barnett, Manly, & Cicchetti, 1993).
2.3.2. Placement history and experience
Living situations and placements are defined differently depending on the federal rules, viewpoints (e.g., youth, agency, or researcher), and other conceptual frameworks being considered. The
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Northwest Alumni Study defined a child's living situation as the place where the youth lived, and
what was viewed by the caseworker as the child's “home”—temporary or not. Following most of the
Federal and CWLA National Data Analysis System work group definitions (Woodruff, 2004), the
study considered the following living situations to be distinct placements:
Initial shelter care
Foster care
Kinship care
Treatment foster care
Group homes
Residential treatment
Independent living placements3
Juvenile justice placements once the youth is under child welfare agency supervision (not as an
initial placement).
Although recorded in the case records, the following living situations were not counted as
Returning to or living with birth parents
Running away from a current placement
Visits to medical hospitals
Respite care
Living with unlicensed friends or relatives4
Unless recorded as the first placement in the log, stays at psychiatric institutions that were fewer
than 30 days (i.e., stays at the beginning of the placement log or 30 days or longer were counted).
• A stay in a juvenile justice or correctional facility that was recorded in the log as the first
2.3.3. Interviews
Professionally trained interviewers from the University of Michigan Survey Research Center
(SRC) administered the interview. In addition to educational attainment and information on the
financial situation of alumni, the interview included questions concerning ethnicity, access and
participation in educational, mental health, and independent living services, resources upon leaving
care, and foster family and other nurturing supports while in care.
2.4. Procedure
The study reviewed case records for the total sample of 659 alumni and interviewed 479 of
these alumni between September 2000 and January 2002. Not every alumnus could be located
During the period covered by this study (1988 to 1998), few of the youth experienced formal supervised apartments
or group home Independent Living (IL) placements. Because the youth were on an active child welfare caseload and were
seen on a periodic basis, however, the researchers decided to include IL situations as placements.
Living with birth family or unlicensed friends or relatives is not counted as an official placement because this study,
and many others, cannot distinguish the nature of these living situations (e.g., the degree to which the child welfare
agency is involved in supervising the child's living situation). Instead of including this condition as a placement, it was
included as one of the variables that described alumni placement history.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 2
Expanded hierarchical type of child maltreatment by birth family of the Northwest alumni a
Child maltreatment by birth family
% (S.E.)
No child maltreatment
Sexual abuse only b
Sexual and other maltreatment (besides emotional maltreatment)
Physical abuse only b
Physical neglect only b
Physical neglect and physical abuse only b
Emotional maltreatment only
Sample size
6.7 (0.4)
4.4 (0.3)
49.2 (0.8)
8.6 (0.5)
10.8 (0.4)
16.7 (0.6)
3.6 (0.3)
Excludes educational neglect.
With or without emotional maltreatment.
and interviewed. The minimum response rate (RR1) as defined by the standard definition manual
of the American Association for Public Opinion Research (2000) was used to determine the
response rate. Using this definition, 26 people (4.0%) of the sample were “ineligible” to participate in the interview. This included deceased alumni and those who were in prisons or mental
institutions during the interview period. After removal of these alumni, the response rate was
75.7% (Pecora et al., 2005). With the exception of a landmark study in New York City (Festinger,
1983) and the more recent prospective cohort alumni studies in the Midwest (Courtney et al.,
2004), this response rate is much higher than those that have been obtained in other longer-term
alumni follow-up studies, which are typically about 60% or less.
2.5. Formation of outcome groups
For some of the foster care experience and outcome variables that are used in the study
optimization analyses, the scores for these variables were trichotomized using theory, the distribution of the data, and practice perspectives about what is more likely to result in positive
3. Results
3.1. Risk factors
In addition to demographic variables, risk factors are variables that may predict alumni outcomes.
Because this study focused on the relationships between outcomes and the experiences while in
foster care that agencies have some influence over (through policies, program approaches, and
resource allocations), demographic and other risk variables were controlled prior to the main analyses. Some of the major risk factors are summarized below.
3.1.1. Types of child maltreatment experienced by the alumni
Child maltreatment is the most frequent reason why children are removed from the birth family
and placed in foster care. Table 2 categorizes maltreatment information recorded in the case records
according to the most prominent types perpetrated by birth family members. Almost all (93.3%) of
the alumni experienced some form of maltreatment by their birth family. Sexual abuse and another
form of maltreatment (physical abuse or physical neglect) was the most frequent type of maltreatment reported for these youth. In this classification, emotional maltreatment was conceptualized
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 3
Risk factors of the Northwest alumni
Risk factors
Reason for initial placement
Other a
Birth parents' substance abuse problems
Child behavior problems
Birth mother
Had substance abuse problems
Had criminal problems
Birth father
Had substance abuse problems
Had criminal problems
Mental/physical health problems diagnosed before or during care
Physical or learning disability
Other impairments b
Sample size
% (S.E.)
64.3 (0.7)
49.2 (0.8)
28.5 (0.6)
19.6 (0.6)
64.6 (1.4)
35.0 (1.4)
45.1 (1.5)
36.7 (1.4)
13.7 (0.5)
13.1 (0.4)
6.4 (0.3)
Includes caregiver(s)' inability or unwillingness to provide care, family stress, and other reasons.
Includes drug exposure, Fetal Alcohol Effects, Fetal Alcohol Syndrome, or visual/hearing impairments.
as commonly occurring with other forms of maltreatment, but as a secondary type if it was the only
from of maltreatment present. Of all types, emotional maltreatment (unaccompanied by any other
form of maltreatment) was experienced least often by alumni.
3.1.2. Reasons for initial placement
Table 3 provides data from case records regarding the reasons for initial placement into foster
care. Because many alumni were placed for multiple reasons, the categories are not mutually
exclusive. The most common reason for initial placement was child maltreatment (64.3%), while
the least common was child behavior problems (19.6%).
3.1.3. Parents' substance abuse and criminal problems
Table 3 also includes information about the alumni family situation before or during foster
care. The highest proportion of parent problems was in the area of substance abuse (64.6% for
mothers and 45.1% for fathers). Meanwhile, slightly over one-third of parents experienced
criminal problems (35.0% of mothers and 36.7% of fathers).
3.1.4. Mental/physical health problems diagnosed before or during foster care
The proportion of the sample that had mental or physical health diagnoses is presented at the
bottom of Table 3. The most frequent diagnosis was ADHD (13.7%).
3.2. Foster care experiences
3.2.1. Placement experiences
Alumni entered foster care on average at age 11.1 (SE = 0.1) and exited, on average, at age 18.5
(S.E. = b0.1). These average ages are older than some studies because Casey served some older
youth as part of a transition program and all of the agencies did not discharge older youth until
they had completed high school.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 4
Placement history and experience of the Northwest alumni
Placement history and experience
Number of placements
Length of time in care
Placement change rate
Number of reunification failures
Number of runaways
Number of unlicensed living situations with friends/relatives
Sample size
% (S.E.)
Low (3 or fewer)
Medium (4 to 7)
High (8 or more)
Low (fewer than 3.6 years)
Medium (3.6 to 5.9 years)
High (5.9 or more years)
Low (fewer than 0.61 placements/year)
Medium (0.62 to 1.23 placements/year)
High (1.23 or more placements/year)
Low (0)
Medium (1)
High (2 or more)
Low (0)
Medium (1)
High (2 or more)
Low (0)
Medium (1)
High (2 or more)
31.9 (0.7)
35.8 (0.7)
32.3 (0.7)
32.5 (0.7)
27.6 (0.7)
39.9 (0.7)
27.6 (0.6)
29.3 (0.7)
43.1 (0.8)
69.5 (0.7)
18.8 (0.7)
11.7 (0.6)
60.7 (0.8)
18.1 (0.6)
21.2 (0.7)
76.8 (0.7)
16.8 (0.6)
6.4 (0.4)
The optimal level of each placement history and experience variable is bolded/italicized. These will be used in
subsequent optimization analyses.
Many youth had unstable living situations while in care: alumni had an average of 6.5
placements (S.E. = 0.1), an average length of time in care of 6.1 years (S.E. b 0.1), and an average
placement change rate of 1.4 placements per year (S.E. b 0.1). Almost one-third experienced eight
or more placements. While most youth did not have any reunification failures (a return home,
followed by re-entry into foster care), over 1 in 10 had two or more failures, and over 1 in 5 had
run away at least twice (see Table 4).
3.2.2. School changes, services and activities while in foster care
In accordance with the placement instability noted above, nearly one-third of the alumni
reported 10 or more school changes from elementary through high school. Some of these changes,
however, may have occurred prior to and/or after discharge from foster care.
The degree of youth access to services while in foster care is a concern of many advocates and
staff members, but over 8 in 10 alumni reported access to “a lot of” therapeutic services during
care. The proportion of alumni who reported that they had participated in fun activities with their
foster family was similar (76.1%) (see Table 5).
3.2.3. Preparation and resources upon leaving foster care
Slightly over half of the alumni reported that they had been somewhat prepared or very prepared
for independent living. But as noted in Table 6, only about one-quarter to one-third reported having
resources when they left care such as a driver's license, cash, or dishes and utensils (see Table 6).
3.2.4. Foster family and other nurturing supports while in care
Over four in five alumni reported feeling loved while in care; over 60% found their foster
parents to be some or a lot helpful; but slightly less than half reported having a close and
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 5
Services and activities while in foster care
Services and activities while in foster care
Education services and experience
Had access to tutoring or other supplemental education services a
Actually participated in tutoring or other supplemental education services
Total number of school changes, elementary to high school
(including any normal transitions from elementary to middle
school, and middle to high school)
Therapeutic services and supports
Had access to counseling and mental health services
Had access to alcohol and drug treatment programs
Had access to group work or group counseling
Had access to a lot of therapeutic services and supports
Activities with foster family
Participated in fun activities with foster family
Participated in religious activities with foster family
Participated in both fun activities and religious activities
(sum of the two activities with foster family items above)
Sample size
% (S.E.)
Low (3 to 6)
Medium (7 to 9)
High (10 or more)
89.1 (1.0)
48.1 (1.5)
35.0 (1.4)
34.8 (1.4)
30.2 (1.4)
95.6 (0.5)
93.9 (0.7)
86.3 (1.0)
83.6 (1.0)
76.1 (1.3)
54.5 (1.5)
45.9 (1.5)
The optimal level of each services and activities variable is bolded/italicized. These will be used in subsequent
optimization analyses.
confiding relationship with an adult while growing up. About one-third of alumni reported
some maltreatment while in care, but these data include alleged as well as substantiated
reports and vary from the case record review findings (17.6% of the alumni were alleged to
have been abused by a foster family member in the case record review). Nationally, rates of
child abuse or neglect among youth placed in foster care vary widely (ranging from less than
1% to 6% or higher), in part due to measurement problems such as what behaviors are
counted and the time period that is used (e.g., Poertner, Bussey & Fluke, 1999; U.S.
Department of Health and Human Services, 2005b). So these data should be viewed with
caution (see Table 7).
3.3. Outcomes and benchmarks: education
3.3.1. Education: high school completion
Despite the fact that 65% of alumni experienced seven or more school changes from
elementary through high school, they completed high school at similar rates (84.8%) as the
general population rate of 87.3% for ages 18 to 29 (U.S. Census Bureau, 2000a, p. 1). GED
completion rates were nearly 30% (see Table 8).
3.3.2. Education: postsecondary enrollment and completion rates
Table 8 also indicates that although two in five alumni received some education beyond
high school, less than half of these (20.6%) completed a degree or certificate. One in six
completed a vocational/technical degree and only 1 in 50 completed a Bachelor's or higher
degree. For alumni ages 25–33, the Bachelor's completion rate (2.7%) was much lower than
for the general population in a similar age range of 25 to 34 years (24.4%) (U.S. Census
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 6
Preparation and resources for leaving foster care
Preparation and resources for leaving foster care
% (S.E.)
Preparation for leaving care
Had access to employment training or job location services
Had access to independent living training groups or workshops
Alumni reporting they were somewhat or very prepared for independent living
Alumni who had health insurance at exit
Degree of preparation for leaving care
Low (0 or 1)
(sum of four preparation for leaving care items above) a
Medium (2)
High (3 or 4)
Resources upon leaving care
Alumnus had the following when leaving care:
Sum of three resources upon leaving care above
Sample size
A driver's license
$250 in cash
Dishes and utensils
Low (0)
Medium (1)
High (2 or 3)
84.2 (1.1)
67.9 (1.3)
56.9 (1.5)
47.4 (1.5)
17.7 (1.1)
24.8 (1.3)
57.5 (1.5)
33.3 (1.4)
38.4 (1.4)
23.7 (1.3)
43.5 (1.5)
27.7 (1.3)
28.8 (1.3)
The optimal level of each preparation and resources for leaving care variable is bolded/italicized. These will be used in
subsequent optimization analyses.
Bureau, 2000b, p. 1). At the time of interview, over 1 in 10 alumni (11.9%) were enrolled in
3.3.3. Outcomes and benchmarks: employment and finances
This section presents data related to four areas of economic well-being: Homelessness,
employment, receipt of public assistance, and personal and household finances.
3.4. Homelessness
More than one in five alumni experienced homelessness for 1 day or more within a year of
leaving foster care (see Table 9). This rate is considerably higher than the rate of homelessness
found in a Wisconsin study of alumni 12 to 18 months after they left foster care (12%) (Courtney,
Piliavin, Grogan-Kaylor, & Nesmith, 2001), and the rate found in a Washington State study of 19to 20-year-old alumni who had been out of foster care for the past 12 or more months (11%)
(Brandford & English, 2004).
3.4.1. Employment and receipt of public assistance
The employment rate among the alumni who were eligible for work was 80.1%, which was
substantially lower than the national average of 95% for ages 20 to 34 during 2000 (see Table 9).
About one in six alumni (16.8%) was receiving cash public assistance from Temporary Aid to Needy
Families (TANF) or General Assistance at the time of their interview. This rate nearly matches that of
a foster care alumni study in Wisconsin (16.6%; see Dworsky, 2005), but was lower than what Casey
Family Services found in a long-term follow-up study (26%; see Casey Family Services, 1999,
p. 13). The Northwest Alumni Study rate, however, is more than five times higher than the general
population rate in 2000 (3% of U.S. households with reported public assistance income in 1999; U.S.
Census Bureau, 2000c, Table DP-3) (see Table 9).
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 7
Foster family and other nurturing supports while in care
Foster family and other nurturing supports
Positive parenting by foster parents
% (S.E.)
Felt loved while in foster care
Overall, foster parents were helpful (rated from 1 to 7)
A little (1 to 4)
Some (5 to 6)
A lot (7)
Had a close and confiding relationship with an adult while growing up
Foster family helped with ethnic identity
Child maltreatment while in foster care
Some child maltreatment
by foster parent or other caregiver a
Sexual abuse only
Sexual abuse and other maltreatment
(physical abuse or physical neglect)
Physical abuse only
Physical neglect only
Physical neglect and physical abuse only
Sample size
51.6 (1.5)
25.6 (1.3)
22.8 (1.2)
81.5 (1.2)
36.0 (1.5)
28.1 (1.4)
35.9 (1.4)
45.7 (1.5)
62.2 (1.5)
32.8 (1.4)
3.7 (0.6)
4.0 (0.5)
5.6 (0.7)
10.1 (0.9)
9.4 (0.9)
The foster care variables for which optimal levels have been specified and used in optimization analyses are bolded/
italicized. The optimal level of child maltreatment was no child maltreatment.
3.4.2. Finances
Many alumni were experiencing difficulty in finding jobs that paid living wages. One-third of
the alumni lived in households that were at or below the poverty line, which was three times the
national poverty rate (U.S. Census Bureau, 2001, pp. 2 and 5). More alumni also lacked health
insurance coverage than the 18- to 44-year-old age group in the general population (33.0% vs.
18%) (Institute of Medicine, 2001, p. 2) (see Table 9).
4. What program changes could improve youth outcomes?
Child welfare administrators and practitioners want to know the aspects of service delivery that
they should improve to maximize success for youth in care. To address this need, this study
conducted statistical simulations that estimated the degree to which optimizing certain foster care
experiences might affect alumni outcomes.
The first step in the simulation was to create a domain score for each alumnus on the
Education and Employment and Finance domains. For example, the Education domain
Table 8
Education outcomes of the Northwest alumni
Education outcomes
% (S.E.)
Completed high school—high school diploma or GED
Completed high school—high school diploma only
Any education past high school (any type of postsecondary education)
Completed any degree/certificate beyond high school (vocational, Bachelor's, etc.)
Completed college or more (Bachelor's or higher)
Sample size
84.8 (1.9)
56.3 (2.7)
42.7 (2.7)
20.6 (1.8)
1.8 (0.4)
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 9
Employment and finances outcomes of the Northwest alumni
Homelessness, public assistance, finance and employment outcomes
% (S.E.)
Homeless within a year of leaving care
22.2 (2.1)
Employment and public assistance
Employment among those eligible to be in the workforce a
Received cash public assistance anytime after age 18
Currently receives cash public assistance (AFDC/TANF/General Assistance) b
Alumni household had received cash public assistance in the past 6 months
80.1 (1.8)
51.7 (2.8)
16.8 (2.2)
47.8 (2.8)
Household income at or below the poverty level
Household income greater than three times the poverty level
Has no health insurance
Owns house or apartment
Sample size
33.2 (2.6)
21.3 (2.2)
33.0 (2.6)
9.3 (1.6)
Alumni who were not in the workforce included students, homemakers, and people with severe disabilities.
AFDC: aid to families with dependent children; TANF: temporary aid for needy families.
outcome score summed the five individual education outcome variables that are listed in Table
8. Scores could range from zero (no positive individual education outcomes) to five (attained
all positive individual education outcomes). The same procedure was used for the nine
Employment and Finance variables listed in Table 9. For both outcome domain scores, the
simulation defined the individual outcome variables so that they were all in a positive direction
(e.g., high school completion, not homeless after foster care, and so on). Therefore, outcome domain
scores represented the number of positive items within Education and Employment and Finance,
The second step in the simulation was to use the two outcome domain scores to create regression
models that would estimate the number of positive outcomes in each domain that each alumnus
would achieve based on the foster care experiences that the alumnus actually had in foster care (e.g.,
time in care, number of schools attended). Next, these foster care experiences were “optimized” in
the simulation so that each alumnus had the most optimal foster care experience possible (e.g., a
short time in care, fewer number of schools attended). The foster care variables and their optimal
levels are bolded and italicized in Tables 4–7.
The last step in the simulation used the optimized variables in a second regression equation to
estimate the number of positive outcomes that could be achieved with these optimal experiences.
The change in the predicted number of positive outcomes before and after optimizing the foster
care experience variables represents the anticipated effect of optimizing these foster care
experiences. (See Little, 1982 for more information on this statistical approach.) This statistical
simulation was conducted using all foster care experience areas at once (to estimate the effects of
optimizing all areas simultaneously), and on each area individually (to measure the effect of
optimizing only a particular area). The latter was done because it is more likely that an agency
could implement program changes for one area than all areas at once.
Table 10 presents the estimated outcome domain scores before and after the simulated
optimizations of the foster care experiences. For example, the Education domain scores before
and after optimization of the foster care experience area of Placement History and Experience
Outcome domains
Estimated outcome score after optimization of foster
score before
history and
services and
services and
Education (highest
possible score: 5)
Employment and Finances 5.2
(highest possible score: 9)
care experience areas (percent change in negative outcomes)
for leaving
Foster family and
other nurturing
2.8 (− 17.8)
2.4 (− 4.2)
2.7 (− 14.6) 2.1 (7.0)
3.0 (− 25.5)
5.4 (− 6.8)
5.4 (− 7.2)
5.2 (− 1.5)
5.3 (−3.0)
5.6 (− 12.2) 5.0 (3.2)
6.2 (− 27.9)
A negative value in parentheses represents a decrease in the number of undesirable outcomes; a positive value represents an increase in the number of undesirable outcomes.
Dashes [–] designate foster care experience areas that were not optimized because none of the individual variables within the areas significantly predicted the outcomes for the two
major outcome domains. Furthermore, although some of the percentages in the parentheses appear negligible, some of the variables within the foster care experience area did
significantly predict the outcomes for the domains. However, when examined in combination with other variables within that area, the effects cancelled each other out (i.e., some of
the variables within the foster care experience area serve to increase positive outcomes, while others decrease positive outcomes.)
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Table 10
Outcome domain scores before and after optimization of foster care experience areas
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
were 2.3 and 2.8, respectively. This represents a 17.8% decrease in the number of undesirable
outcomes. Thus 17.8% is presented in parentheses due to improving placement history and
experience. This statistic is calculated using the following formula:
ðnumber of negative outcomes after optimization−number of negative outcomes before
optimizationÞ=the of negative outcomes before optimization 100
Note that (a) the number of positive outcomes must be subtracted from the highest possible
score to calculate the number of negative outcomes; and (b) the number of positive outcome
scores in the table will vary slightly due to rounding error in the estimated outcome scores before
and after the optimization. But the percent change in negative outcomes score is based on nonrounded outcome scores.
For the Education domain, optimization of two foster care experience areas reduced the
estimated number of undesirable outcomes significantly: Placement History and Experience and
Resources upon Leaving Care (− 17.8% and − 14.6%, respectively). The effect of Foster Family
Activities was modest. Note that, in this analysis with this sample, optimizing Foster Family and
Other Nurturing Supports was estimated to result in a 7.0% increase in undesirable outcomes.
There were three variables within Foster Family and Other Nurturing Supports that were not
conceptualized well, so those variables may have counteracted the effects of the other variables
within that area. The first was parenting style where the authoritative, permissive, authoritarian
and disengaged dimensions (Baumrind, 1995) may not have been well-defined or were interacting with other variables. The second problematic variable was whether the alumnus had
experienced a close and confiding relationship with an adult while growing up. Many alumni had
mixed scores on this variable—possibly because frequent placement change disrupted these
relationships or the way the variable was interpreted. The third variable with mixed results was
maltreatment while in foster care. Because we included unsubstantiated reports as well as
substantiated reports for both foster parents and other caregivers, this variable may have produced
conflicting results. Despite some concerns about three of the variables within this area, a dramatic
reduction in undesirable education outcomes (− 25.5%) was estimated when optimizing all foster
care experience areas simultaneously.
For the Employment and Finances domain, optimizing the Resources upon Leaving Care
foster care experience area reduced the number of undesirable outcomes (− 12.2%). The estimated
effects of optimizing Placement History and Experience were generally modest. Note that, similar
to the Education domain, optimizing Foster Family and Other Nurturing Supports was estimated
to result in a 3.2% increase in undesirable outcomes. When all foster care experience areas were
optimized simultaneously, there was a dramatic estimated reduction in undesirable employment
and financial outcomes (− 27.9%).
5. Discussion
5.1. Summary of results
Children placed in foster care face many educational challenges. Youth experience frequent
school changes, which, along with other factors such as learning disabilities and child maltreatment,
are associated with academic difficulties (Rumberger & Larson, 1998). Despite the challenges of
child maltreatment, placement instability, and other adversities, many alumni demonstrated positive
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
outcomes in the domains of Education and Employment and Finance. Many alumni did, however,
face significant challenges in these areas:
• Foster care alumni obtained a GED instead of a high school diploma at a rate nearly six times the
rate of the general population (28.5% vs. 5%) (National Center on Education Statistics, 2003).
• Completion rates for postsecondary education were low.
• Alumni were in fragile economic situations: one-third reported household incomes that were at
or below the poverty level and one-third had no health insurance. [Data related to outcomes in
the domains of mental health, physical health, relationships, and parenting are reported in
Pecora et al., in press].
While having a GED is more beneficial than not completing high school (Smith, 2003),
research indicates that people who obtain diplomas instead of GEDs are more successful as
adults; for example, GED recipients are more than twice as likely as those completing regular
high school diplomas to not enroll in post-secondary education (Bozick & DeLuca, 2005), and
have lower incomes than those with diplomas (Grubb, 1999).
Many alumni are part of society's large group of marginalized youth ages 16–24 who are neither
employed nor in school: 13% in the general population (; Nelson,
2004); and nearly one-third in the Midwest foster care population (Courtney, Dworsky, Ruth, Keller,
Havlicek, & Bost, 2005). Statistical simulations that estimated the degree to which optimizing
certain foster care experiences might affect alumni outcomes indicated that certain areas of program
improvement should substantially improve outcomes. Note that this assumes that the associations
between placement characteristics and alumni outcomes found in the regression analyses simulations are indeed due to the effects of the placement experiences. A program experiment that
systematically compared conventional child welfare service approaches to minimizing placement
change (and other key placement experiences) with different approaches that achieved lower rates of
placement change would be an essential next step to test these assumptions.
5.2. Recommendations for policy and program design
The recommendations stem from the basic descriptive outcomes of the study, the foster care
experience statistical simulations, and conversations with stakeholders. Stakeholders included alumni
of foster care, foster parents, caseworkers, and agency executives, as well as clinical and policy
specialists from each of the three collaborating organizations and other public child welfare agencies.
5.2.1. Education
Although high school completion rates were quite high, the rates at which alumni completed
high school with a GED were disproportionately higher than the general population.
1. Foster care programs should encourage youth to obtain a high school diploma and not just a
GED. School personnel would also benefit from more training about the challenges that youth
in foster care face, and ways they can advocate for these youth. (See the Endless Dreams and
other curricula at
2. Improve identification and treatment of mental health problems that may act as barriers to
classroom success. This would involve evidence-based treatment of social phobia, depression,
and the sleep and attention problems that accompany Post Traumatic Stress Disorder (PTSD)
found in a number of alumni studies (Courtney et al., 2005).
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
3. Minimize placement change. By optimizing all variables describing Placement History and
Experience, the statistical simulations estimated a 17.8% decrease in negative education
outcomes. If youth do not change homes and schools, there is no need to transfer school records
and the youth are less likely to fall behind. Having fewer placement changes may allow youth in
care to develop better social support networks, which can assist them to find employment and can
serve as a safety net when they encounter financial difficulties. Placement instability, in part, is a
result of poor administrative processes, lack of agency support of foster parents, and behavioral
problems of youth (James, 2004). All of these factors need to be studied and addressed.
4. Provide concrete resources to youth as they leave care. Optimization of Resources upon Leaving
Care predicted a 14.6% decrease in negative education outcomes. It may be that having concrete
resources such as a driver's license, $250 in cash, and dishes and utensils results in more financial
stability, allowing alumni to pursue their education goals. A more plausible explanation is that
these variables described youth who had received many different opportunities to develop skills
for independent living, and had positive relationships with foster parents, agency staff and other
adults, as well as concrete resources.
5. Support better preparation for, access to, and success in postsecondary education programs.
Caseworkers, foster families, and other stakeholders should encourage young people in foster
care to plan for college or vocational school, and support them in being adequately prepared for
higher education and training. Youth will benefit from information about local collegepreparatory programs, such as Gear-Up, TRIO, and Upward Bound, and help with enrollment
in these programs.5
5.2.2. Employment and finances
Many alumni were living in vulnerable financial and housing situations.
1. Overhaul independent living preparation. Alumni varied widely in their level of readiness for
emancipating from foster care (as evidenced by the uneven findings for employment preparation,
life skills preparation, education, and income).
a. Federal and state funds are being spent on a variety of untested life skills training, employment
services, and education supports (U.S. General Accounting Office, 2004). Some child welfare
leaders advocate redirecting these funds to the most promising programs and rigorously
evaluating them (Clark & Davis, 2000; Massinga & Pecora, 2003; Shirk & Stangler, 2004).
b. Every youth in foster care for 6 months or more should be expected to develop a comprehensive transition development plan that includes planning for supportive relationships,
community connections, education, life skills assessment and development, identity formation, housing, employment training and experience, physical health, and mental health
(Casey Family Programs, 2001; Mech, 2003).
c. The transition plans should increase youth access to Individual Development Accounts
(IDAs), special “youth opportunity passports,” and asset accumulation strategies like debit
accounts. (See for example.)
d. Provide youth who are exiting care with a broad foundation in life skills and some concrete
resources. More extensive preparation for independent living and a critical mass of resources
during exit from care will help alumni avoid homelessness, conserve and build their resources,
and compete in the employment market.
See,, www.ed.
gov/print/programs/gearup/index.html, and Casey Family Programs (2003b).
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
e. Foster care services should be extended to age 21 and independent living services to age 25. Or
at a minimum, “booster session” programs could provide a toll-free phone number and various
fallback services to alumni after they leave care. This service could also include ongoing
access to special job and housing search assistance well beyond the current age limitations.
2. Provide targeted education support services and enrichment experiences. By increasing access
to supplemental educational services and tutoring, and by minimizing the number of school
changes, alumni employment and financial outcomes will be improved, along with completion
rates for post-secondary education and training.
3. Strengthen housing programs and other supports to prevent homelessness after leaving care.
a. Youth should be encouraged to develop and maintain lifelong relationships with foster
parents and other supportive adults so alumni have a place to go during difficult times. This
may require aftercare supports for the foster parents to help support this key activity.
b. Transitional housing and public/community housing systems need to be strengthened. Government agencies can work with local Section 8 landlords to help allocate apartments for lowincome foster care alumni. As with other groups whose special needs have been recognized
(such as battered women), alumni would benefit from these new housing models that provide
not only housing subsidies but also home-based case management or other adult guidance
such as Scattered-Site, sober living, Master-Lease Models, and HUD, HOME, and Section
8 housing assistance (Choca et al., 2004; Polcin, 2001; Van Leeuwen, 2004).
5.3. Conclusion
This study found that many youth who were placed in foster care benefited from the protection,
emotional care, and services that they received while they were in care, while some respondents
did not benefit from this experience. It found that the service delivery systems were unable to
prepare some alumni to secure and sustain jobs that pay a living wage with health insurance, and
to help them complete vocational training or education.
Using statistical simulations, certain program areas were identified that, when optimized, can
improve alumni outcomes. Rigorous field trials are important next steps in confirming these
simulated findings. In other words, the prediction simulations would be bolstered by additional
research identifying the types of services and foster care program performance levels that are
linked with positive adult outcomes.
Apart from the research efforts, many program reforms can be made now, as described above. The
statistical simulations revealed the potential power of targeted program improvements. Finally, many
of the needed improvements will be more successful if they are anchored in larger structural and
community-based reforms that involve the public and private sectors, including neighborhood
associations and local businesses.
American Academy of Pediatrics Committee on Early Childhood, Adoption and Dependent Care. (2000). Developmental
issues for young children in foster care. Pediatrics, 106, 1145−1150.
American Association for Public Opinion Research. (2000). Standard definitions: Final dispositions of case codes and
outcome rates for surveys. Ann Arbor, MI: AAPOR
Barnett, D., Manly, J. T., & Cicchetti, D. (1993). Defining child maltreatment: The interface between policy and research.
In D. Cicchetti & S. Toth (Eds.), Child abuse, child development and social policy. Journal in Applied Developmental
Psychology Vol 8. Norwood, NJ: Ablex Publishing Corporation.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Barth, R. P. (1990). On their own: The experience of youth after foster care. Child and Adolescent Social Work, 7(5),
Barth, R. P. (1997). The costs and benefits of adoption. In R. Avery (Ed.), Public adoption policy. New York, NY: Auburn House.
Barth, R.P., Lee, C., Wildfire, J., & Guo, S. (in press). Estimating costs of foster care vs. adoption using propensity score
matching. Social Service Review.
Baumrind, D. (1995). Child maltreatment and optimal caregiving in social contexts (Vol. 1). Michigan State University
series on children, youth, and families.
Berrick, J. D., Needell, B., Barth, R. B., & Johnson-Reid, M. (1998). The tender years: Toward developmentally sensitive
child welfare services for very young children. New York: Oxford University Press.
Biehal, N., & Wade, J. (1996). Looking back, looking forward: Care leavers, families and change. Children and Youth
Services Review, 18, 425−446.
Blome, W. W. (1997). What happens to foster kids: Educational experiences of a random sample of foster care youth and a
matched group of non-foster care youth. Child and Adolescent Social Work Journal, 14(1), 41−53.
Bozick, R., & DeLuca, S. (2005). Better late than never? Delayed enrollment in the high school to college transition. Social
Forces, 84, 527−550.
Braitman, L. E., & Rosenbaum, P. R. (2002). Comparing treatments using comparable groups of patients. Annals of
Internal Medicine, 137(8), 693−695.
Brandford, C., & English, D. (2004). Foster youth transition to independence study. Seattle, WA: Office of Children's
Administration Research, Washington Department of Social and Health Services.
Buehler, C., Orme, J. G., Post, J., & Patterson, D. A. (2000). The long-term correlates of family foster care. Children and
Youth Services Review, 22(8), 595−625.
Casey Family Programs. (2001). It's my life — a framework for youth transitioning from foster care to successful
adulthood. Seattle, WA: Casey Family Programs.
Casey Family Programs. (2003a). Family, community, culture: Roots of permanency — a conceptual framework on
permanency from Casey Family Programs. Seattle, WA: Casey Family Programs.
Casey Family Programs. (2003b). Higher education reform: Incorporating the needs of foster youth. Seattle, WA: Author
Casey Family Services. (1999). The road to independence: Transitioning youth in foster care to independence. Shelton,
CT: Author (
Choca, M. J., Minoff, J., Angene, L., Byrnes, M., Kenneally, L., Norris, D., et al. (2004). Can't do it alone: Housing
collaborations to improve foster youth outcomes. Child Welfare, 83(5), 469−492.
Clark, H. B., & Davis, M. (2000). Transition to adulthood: A resource for assisting young people with emotional or
behavioral difficulties. Baltimore, MD: P.H. Brookes.
Coulling, N. (2000). Definitions of successful education for the ‘looked after’ child: A multi-agency perspective. Support
of Learning, 12(1), 30−36.
Courtney, M. E., Dworsky, A., Ruth, G., Keller, T., Havlicek, J., & Bost, N. (2005). Midwest evaluation of adult functioning
of former foster youth: Outcomes at age 19. Chicago, IL: University of Chicago, Chapin Hall Center for Children.
Courtney, M., Piliavin, I., Grogan-Kaylor, A., & Nesmith, A. (2001). Foster youth transitions to adulthood: A longitudinal
view of youth leaving care. Child Welfare, 80, 685−717.
Courtney, M. E., Terao, S., & Bost, N. (2004). Midwest evaluation of the adult functioning of former foster youth:
Conditions of youth preparing to leave state care in Illinois. Chicago, IL: University of Chicago, Chapin Hall Center for
Cuddeback, G. S. (2004). Kinship family foster care: A methodological and substantive synthesis of research. Children
and Youth Service Review, 26, 623−639.
Curtis, P. A., Grady Jr., D., & Kendall, J. C. (Eds.). (1999). The foster care crisis: Translating research into policy and
practice. Lincoln, NE: University of Nebraska Press.
Dore, M. M., & Eisner, E. (1993). Child-related dimensions of placement stability in treatment foster care. Child and
Adolescent Social Work Journal, 10, 301−317.
Dworsky, A. (2005). The economic self-sufficiency of Wisconsin's former foster youth. Children and Youth Service
Review, 27(10), 1085−1118.
Festinger, T. (1983). No one ever asked us … A postscript to foster care. New York: Columbia University.
Goerge, R., Wulczyn, F., & Fanshel, D. (1994). A foster care research agenda for the 90's. Child Welfare, 73(5), 525−547.
Grubb, W. N. (1999). Learning and earning in the middle: The economic benefits of sub-baccalaureate education. New
York: Community College Research Center.
Holdaway, D. M., & Ray, J. (1992). Attitudes of street kids toward foster care. Child and Adolescent Social Work Journal,
9(4), 307−317.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Institute of Medicine. (2001). Coverage matters: Insurance and health care. Retrieved January 21, 2005, from http://www.
Jacobson, E., & Cockerum, J. (1976, November–December). As foster children see it: Former foster children talk about
foster family care. Children Today, 42, 32−36.
Jaffee, B., & Kline, D. (1970). New payment patterns and the foster parent role. Washington, DC: Child Welfare League of
James, S. (2004). Why do foster care placements disrupt? An investigation of reasons for placement change in foster care.
The Social Service Review, 78(4), 601−627.
Jones, M. A., & Moses, B. (1984). West Virginia's former foster children: Their experiences in care and their lives as
young adults. Washington, DC: Child Welfare League of America.
Kerman, B., Barth, R. P., & Wildfire, J. (2004). Characteristics and costs of post-emancipation support services for children
receiving Casey Family Service. Child Welfare, 83, 239−262.
Kessler, R. C., & Magee, W. J. (1993). Childhood adversities and adult depression. Psychological Medicine, 23, 679−690.
Lindsey, D., & Schwartz, I. M. (2004). Advances in child welfare: Innovations in child protection, adoptions and foster
care. Children and Youth Services Review, 26(11), 999−1005.
Little, R. J. A. (1982). Direct standardization: A tool for teaching linear models for unbalanced data. The American
Statistician, 36, 38−43.
Massinga, R., & Pecora, P. J. (2003). Providing better opportunities for older children in the child welfare system. The
Future of Children, 14(1), 151−173 (
Mech, E. V. (2003). Uncertain futures: Foster youth transition to adulthood. Washington, DC: Child Welfare League of
Minty, B. (1999). Annotation: Outcomes in long-term foster family care. Journal of Child Psychology and Psychiatry and
Allied Disciplines, 40(7), 991−1001.
National Center on Education Statistics. (2003). Digest of Education Statistics, 2001. Table 107. Distribution of 18–29
year olds, by high school completion status and selected characteristics: 1998 to 2000.Washington, DC: Author
(Downloaded August 1, 2003 from
Nelson, D. (2004). Moving youth from risk to opportunity. The 2004 KIDS COUNT Data. Baltimore, MD: Annie E. Casey
Foundation (
Newton, R. R., Litrownik, A. J., & Landsverk, J. A. (2000). Children and youth in foster care: Disentangling the relationship
between problem behaviors and number of placements. Child Abuse and Neglect, 24, 1363−1374.
Park, J. M., Metraux, S., & Culhane, D. P. (2005). Childhood out-of-home placement and dynamics of public shelter
utilization among young homeless adults. Children and Youth Service Review, 27(5), 533−546.
Pecora, P. J., Kessler, R. C., Williams, J., Downs, A. C., English, D., White, J., & O'Brien, K. (in press). What works in
foster care? Oxford, England: Oxford University Press.
Pecora, P. J., Kessler, R. C., Williams, J., O'Brien, K., Downs, A. C., English, D., et al. (2005). Improving family foster
care: Findings from the Northwest Foster Care Alumni Study. Seattle, WA: Casey Family Programs (
Pecora, P. J., Massinga, R., & Mauzerall, H. (1997). Measuring outcome in the changing environment of child welfare
services. Behavioral Healthcare Tomorrow, 6(2), 2−6.
Pew Commission. (2004). Fostering the future: Safety, permanence and well-being for children in foster care. The Pew
Commission on Children in Foster Care. Pew Charitable Trusts. Downloaded June 8, 2004 from http://www.pewtrusts.
Polcin, D. L. (2001). Sober living houses: Potential roles in substance abuse services and suggestions for research.
Substance Use and Misuse, 36(3), 301−311.
Poertner, J., Bussey, M., & Fluke, J. (1999). How safe are out-of-home placements? Children and Youth Services Review,
21(7), 549−563.
Rice, D. L., & McFadden, E. J. (1988). A forum for foster children. Child Welfare, 67(3), 231−243.
Rubin, D. M., Alessandrini, E. A., Feudtner, C., Mandell, D. S., Localio, A. R., & Hadley, T. (2004). Placement stability
and mental health costs for children in foster care. Pediatrics, 113, 1336−1341.
Rumberger, R., & Larson, K. (1998). Student mobility and the increased risk of high school dropout. American Journal of
Education, 107(1), 1−35.
Runyan, D. K., & Gould, C. L. (1988). Foster care for child maltreatment: Impact on delinquent behavior. In G. T Hotaling,
D. Finkelhor, J. T. Kirkpatrick, & M. A. Straus (Eds.), Coping with family violence: Research and policy perspectives
(pp. 224−237). Thousand Oaks, CA: Sage Publications, Inc.
Ryan, J., & Testa, M. (2004). Child maltreatment and juvenile delinquency: Investigating the role of placement and
placement instability. Champaign-Urbana, IL: University of Illinois at Urbana-Champaign School of Social Work,
Children and Family Research Center.
P.J. Pecora et al. / Children and Youth Services Review 28 (2006) 1459–1481
Ryan, J., & Testa, M. (2005). Child maltreatment and juvenile delinquency: Investigating the role of placement and
placement instability. Children and Youth Services Review, 27, 227−249.
Scarcella, C. A., Bess, R., Zielewski, E. H., Warner, L., & Geen, R. (2004). The cost of protecting vulnerable children IV.
Washington, DC: Urban Institute.
Sedlak, A. J., & Broadhurst, D. D. (1993). Study of adoption assistance impact and outcomes: Final report. Report to the
U.S. Department of Health and Human Services, Westat, Rockville, MD.
Shirk, M., & Stangler, G. (2004). On their own: What happens to kids when they age out of the foster care systems.
Boulder, CO: Westview Press.
Smith, D. K., Stormshak, E., Chamberlain, P., & Whaley, R. B. (2001). Placement disruption in treatment foster care.
Journal of Emotional and Behavioral Disorders, 9(3), 200−205.
Smith, T. M. (2003). Who values the GED? An examination of the paradox underlying the demand for the general
educational development credential. Teacher's College Record, 105(3), 375−415.
U.S. Census Bureau. (2000). Percent of high school and college graduates of the population 15 years and over, by age,
sex, race, and Hispanic origin: March 2000. (Table 1a).Washington, D.C.: Author (Downloaded February 10, 2005,
U.S. Census Bureau. (2000b). Educational attainment by sex: 2000 (Table QT-P20.) http://fact?
U.S. Census Bureau. (2000). Profile of selected economic characteristics: 2000. (March 2001 Current Population Survey:
Table DP-3.). Washington, DC: Author.
U.S. Census Bureau. (2001). Poverty status of people by selected characteristics in 2000. Current Population Survey,
March 2001.Washington, DC: Author (
U.S. Department of Health and Human Services, Children's Bureau (2003). Child welfare outcomes 2001: Annual report
to Congress.Washington, DC: U.S. Government Printing Office (
U.S. Department of Health and Human Services, Administration for Children and Families, Children's Bureau. (2005).
Findings from the initial child and family service reviews 2001–2004. Washington DC: Author.
U.S. Department of Health and Human Services, Administration for Children and Families, Children's Bureau. (2005).
The AFCARS report: Preliminary FY 2003 estimates as of August 2005.Washington, DC: Author (Downloaded August
15, 2005 from
U.S. Department of Health and Human Services, Administration for Children and Families, Children's Bureau. (2005).
Trends in foster care and adoption—FY-2000 to FY-2004.Washington, DC: Author (Downloaded December 15, 2005
U.S. Department of Health and Human Services, Administration for Children and Families, U.S. Children's Bureau. (2001).
National survey of child and adolescent well-being: State child welfare agency survey: Report. Washington, DC: Author.
U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, Center of
Mental Health Services. (1999). Consumer-oriented mental health report card: The final report of the Mental Health
Statistics Improvement Program (MHSIP) Task Force on a consumer-oriented mental health report card. Washington, DC:
U.S. General Accounting Office. (2004). Foster youth: HHS actions could improve coordination of services and
monitoring of states' independent living programs. Washington, DC: Author (Report No. GAO-05-25).
Van Leeuwen, J. (2004). Reaching the hard to reach: Innovative housing for homeless youth through strategic partnerships.
Child Welfare, 83(5), 453−468.
Wald, M. S., Carlsmith, J. M., & Leiderman, P. H. (1988). Protecting abused and neglected children. Palo Alto, CA:
Stanford University Press.
Wilson, K. (2006). Foster family care in the United Kingdom. In C. McCauley, P. J. Pecora, & W. E. Rose (Eds.),
Enhancing the well-being of children and families through effective interventions: International evidence for practice
London: Jessica-Kingsley Publishers.
Woodruff, K. (2004). Placement stability definitions to promote consistency in state data reporting for the federal
outcome measure. Washington, DC: Child Welfare League of America National Working Group to Improve Child
Welfare Data.
Wulczyn, F., Barth, R. P., Yuan, Y. Y., Jones-Harden, B., & Landsverk, J. (2005). Beyond common sense: Child welfare,
child well-being, and the evidence for policy reform. Somerset, NJ: Transaction Aldine.
Wulczyn, F. H., & Goerge, R. M. (1992). Foster care in New York and Illinois: The challenge of rapid change. Social
Service Review, 66(2), 278−294.
Young, D. W., & Brandt, A. (1977). Benefit–cost analysis in the social services: The example of adoption reimbursement.
Social Service Review, 51(2), 249−264.