Instability in the lives of foster and nonfoster youth: Mental health impediments and attachment insecurities

Children and Youth Services Review 84 (2018) 159–167

Contents lists available at ScienceDirect

Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth

Instability in the lives of foster and nonfoster youth: Mental health impediments and attachment insecurities☆

T

⁎

Daniel Bederian-Gardnera, Sue D. Hobbsb, Christin M. Oglec, Gail S. Goodmana, , Ingrid M. Cordóna, Sarah Bakanoskya, Rachel Narrd, Yoojin Chaee, Jia Y. Chongf, the NYTD/CYTD Research Groupa a

University of California, Davis, USA California State University, Sacramento, USA c Duke University, USA d University of Virginia, USA e Texas Tech University, USA f University of Illinois, Urbana Champaign, USA b

A R T I C L E I N F O

A B S T R A C T

Keywords: Foster youth Residential instability School instability Mental health Attachment

Foster youth are at risk for negative mental health and psychosocial outcomes, including when they are on the brink of emancipation from care into self-sustained adulthood. Factors believed to affect outcomes among foster youth include residential and school instability. Although frequent moves to new homes and schools are common for youth living in poverty, instability for foster youth involves not only changing homes and schools but often also changes in caregivers, thus putting foster youth at risk for disrupted attachment relationships. For the current study, structural equation models examined links between instability, mental health problems, and attachment insecurities in foster and at-risk nonfoster youth. A model containing instability provided a better fit to the data than a model containing foster care status only. Group comparisons revealed that instability was associated with posttraumatic stress disorder symptoms for foster but not nonfoster youth. Implications of instability in the lives of foster youth are discussed.

1. Introduction Children in foster care are at risk for a variety of negative mental health outcomes, including prior to and after transitioning out of care. These include symptoms of posttraumatic stress disorder (PTSD) and problems with emotional functioning, such as increased depression and anxiety (Courtney & Dworsky, 2006; Courtney, Piliavin, Grogan-Kaylor, & Nesmith, 2001; Havlicek, Garcia, & Smith, 2013). Compared to youth from the general population with similar demographics, foster youth experience significantly higher lifetime rates of PTSD symptomology (Pecora, White, Jackson, & Wiggins, 2009). Many foster youth have suffered child abuse and/or neglect (Child Welfare Information Gateway, 2013; U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children,

Youth and Families, Children’s Bureau, 2016), and such experiences can result in PTSD symptoms (Salazar, Keller, Gowen, & Courtney, 2013). Removal from home and subsequent placement moves can also be traumatic for foster youth (Chambers et al., 2017; Unrau, Seita, & Putney, 2008), contributing to symptoms of PTSD. In addition to foster youth's elevated risk of negative mental health outcomes, foster youth are also at risk of developing insecure attachment with close others (Bowlby, 1980; Schofield & Beek, 2005), which may further increase their vulnerability to mental health problems (Bowlby, 1977; Conradi & Jonge, 2009). According to attachment theory, during distressing or anxiety-producing events, secure youth should exhibit proximity seeking to a trusted individual (Bowlby, 1977), whereas insecure youth are less likely to exhibit such behaviors (Mikulincer, Shaver, & Solomon, 2015). For many youth who are

☆ We thank the California Department of Social Services (CDSS) for providing data for this study. Administrative data were made available through the National Data Archive on Child Abuse and Neglect, Cornell University, Ithaca, New York. The National Youth in Transition Database (NYTD) survey for California (Phase 1) was conducted by the authors through Contract 09-2041 (Goodman, PI) between CDSS and the Regents of the University of California (UC). Nonfoster youth data collection was funded by an internal grant from UC Davis. Neither funding agencies, nor the National Data Archive on Child Abuse and Neglect, bear any responsibility for the analyses or interpretations presented herein. The opinions and conclusions expressed are solely those of the authors and should not be considered as representing the policy of the collaborating agencies or any agency of the California government. ⁎ Corresponding author at: Department of Psychology, University of California, 1 Shields Avenue, Davis, CA 95616, USA. E-mail address: [email protected] (G.S. Goodman).

https://doi.org/10.1016/j.childyouth.2017.10.019 Received 20 June 2017; Received in revised form 16 October 2017; Accepted 16 October 2017 Available online 18 October 2017 0190-7409/ © 2017 Published by Elsevier Ltd.

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

are frequent. Instability among at-risk nonfoster youth and among foster youth also involves changing schools. However, foster youth typically experience the added stress of school instability more frequently than nonfoster youth (Blome, 1997). Repeated school changes can be harmful to youth well-being. It is estimated that every time a foster youth changes schools, it takes 4 to 6 months to recover academically (Casey Family Programs, 2007). Moreover, for foster youth, not only are relationships with attachment figures, such as caregivers, severed, but attachment relationships with teachers and peers are disrupted, leaving youth more vulnerable. Thus, school instability is widely considered a source of distress for foster youth (e.g., Hango, 2006). Conjoined, residential and school instability result in accumulation of risk (Sameroff & Seifer, 1990). Thus combined residential and school instability may be associated with more severe adverse mental health outcomes and attachment insecurities (e.g., Leonard, Stiles, & Gudiño, 2016) compared to residential or school instability alone.

nearing transition out of foster care (often a distressing and/or anxietyproducing event), not only are their existing attachments insecure, but also their access to social support and mental health resources will soon become even more limited than they have been (Courtney et al., 2001; Courtney & Dworsky, 2006; Cunningham & Diversi, 2012), placing them at further risk (Reilly, 2003). During foster care, both negative mental health outcomes and insecure attachments may be further amplified by frequent placement changes that result in residential and school instability (Dolan, Casanueva, Smith, & Ringeisen, 2013; Festinger, 1983; Rubin, O'Reilly, Luan, & Localio, 2007; but see Berger, Bruch, Johnson, James, & Rubin, 2009). Instability in the form of frequent placement changes makes it difficult to form secure attachments with foster parents, and as youth mature into adolescence, with peers (Farineau, Wojciak, & McWey, 2013). Moreover, instability is associated with low self-esteem, delays in school performance, and reduced life satisfaction (e.g., Fernandez, 2007; Reilly, 2003). Even children who enter foster care with low or no known internalizing and externalizing problems show an increase in total problem behaviors in direct correlation with their number of subsequent placements (Newton, Litrownik, & Landsverk, 2000). The purpose of the current study was to gain a better understanding of the possible role of instability in predicting mental health problems and attachment insecurities in foster youth. Our measure of instability combined both residential and school changes. We chose residential changes to reflect instability in homes, families, caregivers, and neighborhoods, whereas school changes captured disruption in peer and teacher relationships. As such, we evaluated relations among foster care status (foster youth vs. at-risk age-matched nonfoster youth), instability, mental health problems, and attachment insecurities in the hope of gaining insight into predictors of negative outcomes for foster youth.

1.2. Attachment across the life course According to attachment theory (Bowlby, 1969, 1988), individuals develop systematic patterns of expectations, beliefs, and emotions concerning the availability and responsiveness of close others in the context of early experiences with caregivers. Over time these beliefs and expectations become internalized and form mental representations (i.e., internal working models) of the self and others in close relationships. Research indicates that individual differences in internal working models influence emotion regulation as well as individuals' perceptions of and coping with stressors and negative life events (e.g., Mikulincer & Shaver, 2016). Individuals who score high on one underlying dimension of attachment, attachment anxiety, experience high levels of distress when caregivers or close others are unavailable or unresponsive. This can be characterized by persistent fear of rejection or abandonment (Bartholomew & Horowotz, 1991). Anxious attachment in adulthood has been linked to a wide range of maladaptive outcomes, including PTSD and depression (Mikulincer, Florian, & Weller, 1993; Ogle, Rubin, & Siegler, 2015). Individuals who score high on the second critical dimension of attachment, attachment avoidance, experience discomfort with close interpersonal relationships and tend to be chronically self-reliant (e.g., Fraley & Shaver, 1997). Attachment avoidance is associated with a broad range of negative outcomes, including reduced social support seeking (Simpson, Rholes, & Nelligan, 1992), affect regulation problems (Mikulincer, Shaver, & Pereg, 2003), increased depression (Wei, Russell, Mallinckrodt, & Vogel, 2007), and greater PTSD symptomology (Fraley, Fazzari, Bonanno, & Dekel, 2006). In contrast to individuals with anxious and avoidant attachment, individuals with secure attachment, as indexed by low levels of attachment anxiety and avoidance, appraise stressful events as less threatening (Mikulincer & Florian, 1998) and seek social support (Mikulincer et al., 1993), which enable them to cope more effectively with stress (Solomon, Ginzburg, Mikulincer, Neria, & Ohry, 1998). Although attachment orientations that are established in childhood theoretically influence the developmental course of emotion regulation and behavior in close relationships throughout adulthood (Bowlby, 1988; Fraley & Shaver, 2000; Hazan & Shaver, 1987), research on the longitudinal stability of attachment indicates that individuals' attachment orientations can be modified by later experiences. In particular, the attachment security of children's bonds with their primary caregivers may decline following negative changes in the family environment and exposure to adverse life events, such as divorce, child maltreatment, and removal from the home into foster care (Belsky & Fearon, 2008; Lewis, Feiring, & Rosenthal, 2000; Thompson, 2013). Moreover, changes in school interrupt positive relationships with teachers, who can also promote secure attachment (e.g., Bergin & Bergin, 2009; Howes & Ritchie, 1999). Compared to youth in normative

1.1. Instability Family instability is associated with adverse outcomes for children over and above negative life events (Marcynyszyn, Evans, & Eckenrode, 2008). Residential and school instability are two key contributions to such findings. Instability among child and adolescent populations is linked to a range of negative mental health and educational outcomes (Jelleyman & Spencer, 2008; Keller, Cusick, & Courtney, 2007). Although frequent relocation is a common experience for many at-risk populations, including economically disadvantaged segments of society, instability is an especially common problem for foster youth. According to the National Survey of Child and Adolescent Well-Being (NSCAW), more than 25% of foster children experienced placement disruptions within their first 18 months in care (Dolan et al., 2013). Foster children report that frequent moves have a substantial negative effect on their mental health (Ellermann, 2007; Keller et al., 2007). Conversely, stable placement is associated with better outcomes: For example, interviews with social workers indicate that children's expressions of unhappiness, nervousness, anxiety, and worry decrease over time when they are in a stable placement (Barber & Delfabbro, 2005). Although selection bias likely influences an understanding of foster youth's reactions to placement instability (Berger et al., 2009), a “cascading effect” may nevertheless still be operative in which mental health problems at first placement contribute to instability, and the instability exacerbates these existing mental health problems and possibly creates even more. A fundamental concern regarding instability for children in foster care is that moving is frequently accompanied by a change in caregiver and severed ties with possible attachment figures (Fernandez, 2007). A change in caregiver causes added stress that is not always experienced by nonfoster youth. According to Bowlby (1980), when children lose their primary caregivers, it is imperative to their mental health to form secure attachments with alternate caregivers. Forming secure attachments to new caregivers can be difficult when placement disruptions 160

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

1. Foster care (compared to nonfoster care) status was expected to be associated with higher levels of insecure attachment, more symptoms of PTSD and more depression/anxiety problems. This hypothesis was tested in the Foster Care Status Model (Fig. 1). 2. Greater instability as measured by a higher number of homes lived in and schools attended was expected to predict higher levels of insecure attachment as well as more symptoms of PTSD and greater depression/anxiety problems in both foster care and nonfoster care youth. This hypothesis was tested in the Instability Model (Fig. 1). 3. Given the critical roles that residential and school stability play in children and adolescents' psychological health and interpersonal relationships, the Instability Model was expected to be a better fit to the data than the Foster Care Status Model. 4. Furthermore, greater instability was expected to be associated with increased mental health problems (e.g., more PTSD and depression/ anxiety symptomology) especially for foster care youth compared to nonfoster youth.

samples, at-risk youth are more likely to exhibit discontinuous attachment representations from childhood to adulthood (Waters, Merrick, Treboux, Crowell, & Albersheim, 2000; Weinfield, Sroufe, & Egeland, 2000). In late adolescence, attachment figures often change from caregivers to close others, such as romantic partners and close friends (Mikulincer et al., 2003). Thus, foster youth who move from home to home and school to school after removal to foster care may be more likely to develop insecure attachment over time compared to their peers raised in more stable environments.

1.3. The current study In the present study, we examined the relation between instability (defined in terms of combined number of homes lived in and number of schools attended) and poor mental health outcomes (i.e., PTSD symptoms, depression/anxiety problems, and attachment insecurities) among 17-year-old foster youth preparing to emancipate into adulthood and 17-year-old youth with no experience in foster care. Data for foster youth were drawn from a western state that participated in the nationwide federally mandated survey of foster youth called the National Youth in Transition Database (NYTD; Chafee, 2008). Data collection for NYTD began in 2010 as commissioned by the Foster Care Independence Act of 1999, which requires all states in the U.S. to examine the longitudinal outcomes of foster youth who age out of care. Foster youth completed a set of mental health and attachment questionnaires that we added to Phase 1 of the State's NYTD survey. The comparison group consisted of nonfoster youth recruited from high schools throughout the western state that serve low socioeconomic status (SES) at-risk students. Based on the aforementioned literature, several hypotheses were advanced.

2. Method 2.1. Participants 2.1.1. Foster youth Participants included 146 foster care youth who completed mental health and attachment questionnaires that were added to Phase 1 of the State's NYTD survey. Foster youth were between the ages of 17 years, 0 months and 17 years 2 months (M = 17 years, 0.29 months, SD = 0.48 months), and approximately 37% were female. Known ethnicities for the foster youth were as follows: 34% Hispanic or Latino/a, 19% African American, 13% Caucasian, 1% Asian, and 33% “Other” or not indicated. Gender and ethnicity data were missing for 30% of foster Fig. 1. Two theorized models: Foster Care Status Model (dashed lines) and Instability Model (solid lines).

161

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

retest reliability (r = 0.90). Based on the LASC data provided by the current sample of youth, coefficient alpha was 0.94.

youth. Youth reported having lived in foster homes (n = 103), in group homes (n = 49), with relatives (n = 59), or in other placements (n = 15). Almost 50% of the foster youth reported having lived in more than one type of placement (n = 66).

2.2.5. California Mental Health Information Survey (CHIS), adolescent version Select questions from the adolescent version of the CHIS were included. The CHIS has been conducted in California since 2001. Details on the methodology and response rates for the 2007 survey are available on the CHIS website: http://healthpolicy.ucla.edu/chis/design/ Pages/methodology.aspx. For the adolescent version of the CHIS, a random sample of 12- to 17-year-old California residents are regularly surveyed about a variety of health topics. Questions are answered on a 5-point scale (1 = all of the time, 5 = none of the time). The present study included the depression/anxiety questions from the CHIS. The section was comprised of 6 items (e.g., “In the last 30 days: How often did you feel hopeless?” “How often did you feel nervous”). Responses were reverse scored so that in the analyses, higher scores represent more severe depression/anxiety problems (0 = none of the time, 4 = all of the time). “Don't know” and “decline to answer” responses were treated as missing. In the current analysis, scores for the six questions were averaged (coefficient alpha = 0.80).

2.1.2. Nonfoster youth The comparison group consisted of 83 nonfoster youth recruited from seven high schools throughout the state that serve low SES at-risk students. We targeted low SES neighborhoods as determined by zip code median yearly income. The mean of zip code median yearly income was $35,693 (SD = $9182.80). The schools spanned five counties across 742 miles throughout the state. These counties serve 43% of foster youth. Nonfoster youth were between the ages of 17 years 0 months and 17 years 11 months (M = 17 years, 4.90 months, SD = 3.30 months), and 59% were female. Nonfoster youth ethnicities were: 55% Hispanic or Latino/a, 6% African American, 7% Caucasian, 11% Asian, and 21% “Other” or no ethnicity indicated. 2.2. Materials 2.2.1. Demographic information For foster youth, demographic information was obtained from administrative records. For nonfoster youth, a standard Demographic Questionnaire was constructed to query age, gender, and race/ethnicity.

2.2.6. Experiences in Close Relationships-Short Form (ECR-S) The ECR-S is a 12-item scale that measures attachment on two dimensions, attachment avoidance (6 items) and attachment anxiety (6 items; Wei et al., 2007). It is a reduced version of the 36-item ECR, which was originally written to be answered in relation to respondents' romantic partners (Brennan, Clark, & Shaver, 1998). The ECR was then revised to be answered in relation to close relationships (e.g., close friends and family members). Thus, for the present study, respondents were asked to answer the ECR-S questions regarding how they feel in close relationships with romantic partners, close friends, or family members. Respondents provide ratings using 7-point scales (1 = strongly agree; 7 = strongly disagree). An example of an avoidant attachment question is, “I want to get close to others, but I keep pulling back.” An example of an anxious attachment question is, “I need a lot of reassurance that I am loved by others.” Coefficient alphas for the ECR-S scales as reported by Wei et al. (2007) are relatively high: α = 0.78 to 0.88 (avoidant attachment), α = 0.77 to 0.86 (anxious attachment). For the present sample, the alphas were somewhat lower than in the Wei et al. study, but in the acceptable range for 6-item scales: α = 0.66 (avoidant attachment), and α = 0.64 (anxious attachment).

2.2.2. NYTD survey The Administration on Children, Youth, and Families (ACYF) of the U.S. Department of Health and Human Services developed a questionnaire for Phase 1 of the NYTD project. These federally required questions can be found in the Federal Register 45 CFR Part 1356, Appendices A and B (Chafee, 2008). Responses to several mandated questions (i.e., regarding homelessness, treatment for drug or alcohol usage, confinement to jail or prison, and having fathered or birthed a living child) were analyzed in the present study to examine the representativeness of our foster care sample. With permission from the State government, additional mental health and attachment questionnaires, and measures of instability, were included at the end of the required questions for the foster youth during the first 2 1/2 months of Phase 1. These measures are described next. 2.2.3. Instability Added to the required questions was one query on the number of homes lived in: “How many homes have you lived in? (A change in your home could be due to moving to a new area with your parents, moving into a group home, moving into an institution, such as the [state] Youth Authority, moving to a new foster home, and so forth. If due to joint custody, you switched back and forth between two parents' homes, this would count as two homes.)” Because we were interested, in large part, in disruption of the child-caregiver relationship associated with residential instability, we worded the question such that joint custody transfers between parents were limited to count as two moves. Added also was one question on the number of schools attended: “How many schools have you attended (grades kindergarten to 12)?” Responses to residential and school instability questions were positively correlated, r = 0.49, p < 0.001, and summed to create a measure of combined residential and school instability.

2.3. Procedure Because the foster youth survey was part of a federally mandated evaluation, parental consent for the foster youth was waived by the State and University Institutional Review Boards. For the nonfoster youth, parental consent was obtained. Nonfoster youth who returned the signed parental consent forms were contacted to complete the survey. Assent was provided by each youth (foster and nonfoster) before the questionnaires were administered. Contact information for foster youth, along with information on youth age, gender, and ethnicity, was procured from the State government. Federal requirements mandated that foster youth be surveyed within 45 days after their 17th birthdays. Contact with the youth within the time limit was attempted but not always successful. Youth who agreed to participate in the study were either administered the survey by telephone, mail, or online. Foster youth who completed the survey received a $5 gift card and were entered into one of three drawings to win $500. All foster youth participants were entered into a final drawing to win $1000. Nonfoster youth were also interviewed by phone, mail, or online. To match the age range of the foster care youth, efforts were made to survey the nonfoster youth within 45 days of their 17th birthdays. Nonfoster youth who completed the survey were given $5 gift cards and

2.2.4. Los Angeles Symptom Checklist (LASC) The LASC (King, King, Leskin, & Foy, 1995) is a self-report measure of overall global distress, PTSD symptomology severity, and PTSD symptoms on three subscales (hyperarousal, avoidance/numbing, and re-experiencing). Questions are answered on a 5-point scale (0 = not a problem, 4 = serious problem). Average scores on PTSD symptomology were analyzed in the current study. The scale has strong psychometric properties including high internal consistency (alpha = 0.95) and test162

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

one of the two mental health measures (scoring over three standard deviations above the mean on either the PTSD or depression/anxiety measure). One nonfoster youth was a multivariate outlier on both measures of mental health. The outliers resulted in violations of univariate and multivariate normality. Thus the scores for these four participants were converted to the next closest extreme scores that were within three standard deviations of their means (Kline, 2011). Age of the youth was tightly constrained to 17-year-olds; therefore age was not considered in the analyses. Descriptive statistics and correlations for key variables are presented in Table 1. Gender (1 = male; 2 = female) was significantly correlated with foster care status (0 = never in foster care; 1 = currently in foster care; r = −0.22, p = 0.003), although not with other key variables. When gender and a path from gender to foster care status were added to the SEM models described below, the model fits did not improve; therefore, gender was not considered further. Ethnicity dummy coded as Hispanic versus nonHispanic did not significantly correlate with the key variables. When ethnicity was instead dummy coded for Caucasian (nonHispanic) (compared to all others) or for African American (compared to all others), the variables were highly skewed, precluding valid analysis. Given the presence of missing data for gender and ethnicity, these variables were not considered further. To test whether foster youth who completed the additional mental health and attachment questionnaires were well-adjusted compared to foster youth who were unwilling to answer the additional questions, we compared the two groups on other negative outcome measures from the federally required questions (i.e., homelessness, substance abuse referral, incarceration, and teen parenthood). Because all four of these outcome variables were categorical, we conducted chi square analyses, which did not reveal significant differences on any of the outcomes between youth who completed the additional questionnaires and youth who did not complete additional questionnaires during the same time period: χ2s (1) ≤ 2.78, ps ≥ 0.10.

were entered into drawings to win $500. 2.4. Plan of analyses Structural equation modeling, using AMOS version 22, was utilized to evaluate the Foster Care Status Model and the Instability Model (Fig. 1 represents both models). AMOS relies on full information maximum likelihood (FIML) to estimate the parameters (Arbuckle & Wothke, 1999). FIML is considered less biased than other methods of dealing with missing data such as listwise and pairwise deletion and mean-imputation (Wothke, 2000). To determine which model provided the best fit to the data, fit indices were compared across the two models. As recommended by Kline (2011), several model fit indices were used: chi-square test of model fit; comparative fit index (CFI; Bentler, 1990); Tucker-Lewis index (TLI; Tucker & Lewis, 1973); and the root mean square error of approximation (RMSEA; Browne & Cudeck, 1992). Good model fit was reported by a nonsignificant chi square, a CFI and TLI > 0.95, and RMSEA < 0.06 (Hu & Bentler, 1999). We compared the two models with cross-validation tests, specifically, the expected cross-validation index (ECVI) and the Akaike information criterion (AIC), which are suitable for comparing nonnested models (Browne & Cudeck, 1992; Kumar & Sharma, 1999). To investigate the possible effect of instability on outcomes across foster care status (foster vs. nonfoster groups), we conducted multiple group analyses of the Instability Model (Fig. 2) dividing the groups into youth in foster care (n = 146) and youth not in foster care (n = 83). 3. Results 3.1. Preliminary analysis We first examined the data for univariate and multivariate outliers. One nonfoster youth and two foster youth were univariate outliers on

Fig. 2. Group model. The foster care variable is removed from the model because it forms the basis of the groups. Paths from instability to attachment avoidance and to attachment anxiety were constrained to zero.

163

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

Table 1 Descriptive statistics and correlations. Variables

Foster youth M (SD)

1. Foster care status 2. Gender 3. Instability 4. Attachment avoidance 5. Attachment anxiety 6.Depression/anxiety 7. PTSD symptoms

Nonfoster youth M (SD)

13.00 (7.29) 3.16 (1.17) 3.43 (1.17) 0.91 (0.75) 0.79 (0.69)

7.88 3.19 3.62 0.83 0.73

1

2

3

4

5

6

7

1

—a 1

0.36⁎⁎⁎ −0.13 1

− 0.01 0.09 0.10 1

− 0.08 − 0.05 − 0.03 0.17⁎ 1

0.05 −0.03 0.14 0.28⁎⁎⁎ 0.40⁎⁎⁎ 1

0.05 0.02 0.25⁎⁎⁎ 0.38⁎⁎⁎ 0.36⁎⁎⁎ 0.60⁎⁎⁎ 1

(4.32) (1.01) (1.16) (0.64) (0.64)

⁎ p < 0.05, ⁎⁎p < 0.01, ⁎⁎⁎p < 0.001. Note. Foster youth: Ns = 85–141; Nonfoster youth: Ns = 69–83. Instability = total number of homes lived in and number of schools attended. Foster care status (0 = nonfoster youth, 1 = foster youth). Gender (1 = male, 2 = female). All remaining variables (higher scores = higher values on each dimension). a χ2 (1) = 7.52⁎⁎.

Table 2 Model fit indices and model comparison. Model fit

Good model fit indices Foster Care Status Model Instability Model

Model comparison

χ2

df

CFI

TLI

RMSEA (90% CIs)

ECVI

AIC

AICC

44.06 1.52

5 4

> 0.95 0.79 1.00

> 0.95 0.10 1.07

≤ 0.05 0.19 (0.14–0.24) < 0.001 (0.00–0.06)

0.39 0.21

88.06 47.52

117.98 78.80

Note. CFI = comparative fit index; TLI = Tucker-Lewis index, RMSEA = root mean square error of approximation; ECVI = expected cross-validation index; AIC = Akaike information criterion; AICC = corrected Akaike information criterion. ECVI and AIC are comparative measures of fit. For χ2, ECVI, and AIC, lower values are preferred.

(β = 0.14, p = 0.05). Instability did not significantly predict attachment anxiety or attachment avoidance scores (p = 0.60 and p = 0.23, respectively, n.s.). As instability in our model was measured by combining number of moves with number of schools, to determine whether one of those variables was driving the results, we examined the paths when the combined instability variable was replaced with either residential instability alone or school instability alone. Paths from instability to PTSD symptoms remained significant when either residential instability (p < 0.001) or school instability (p < 0.01) were entered separately; however, neither residential instability nor school instability was a significant predictor of depression/anxiety scores (p = 0.19, and p = 0.07, respectively, n.s.).

3.2. Model comparison The result of the SEM test indicated that the proposed Instability Model resulted in good fit, whereas the Foster Care Status Model did not. The ECVI, AIC, and AICC confirmed that our hypothesized Instability Model was superior to the Foster Care Status Model. Table 2 presents fit indices and model comparison tests. 3.3. Model paths In the Foster Care Status Model, contrary to our prediction, foster care status did not predict either mental health outcome or attachment scores. Table 3 presents path estimates. In the Instability Model, foster care status significantly predicted greater instability (β = 0.36, p < 0.001). Greater instability in turn significantly predicted more PTSD symptoms (β = 0.24, p < 0.001). There was a trend for greater instability to predict higher levels of depression/anxiety problems

3.4. Group comparison As the paths between instability and both attachment outcomes were nonsignificant in the Instability Model, the paths between these variables were constrained to 0 for the multiple group analyses of the Instability Model. In partial support of our fourth hypothesis, the analyses revealed that the paths between instability and both measures of mental health (PTSD and depression/anxiety) at the group level were significant for youth in foster care (β = 0.28, p < 0.001), but not for youth in the nonfoster group (β = − 0.02, p = 0.83): The z tests comparing coefficients revealed a significant difference between the group paths (CRs > | 1.96|, p < 0.05). Increased instability significantly predicted increased PTSD symptomology for youth in foster care. Greater instability did not predict worse depression/anxiety for either group although the path coefficient approached significance for youth in foster care (β = 0.15, p = 0.07), but not for the nonfoster youth (β = − 0.06, p = 0.55), and z tests comparing the path coefficients failed to find a significant difference between the groups (CRs < | 1.96 |, p > 0.05). Table 4 depicts all path estimates for the two groups.

Table 3 Path estimates for models. Foster Care Status Model

Foster care status → attachment avoidance Foster care status → attachment anxiety Foster care status → depression/anxiety Foster care status → PTSD Foster care status → instability Instability → attachment avoidance Instability → attachment anxiety Instability → depression/ anxiety Instability → PTSD Note. ⁎p = 0.05,

⁎⁎

p ≤ 0.01,

B (S.E.)

β

− 0.03 (0.16) − 0.17 (0.17) 0.43 (0.61)

− 0.01

0.06 (0.09)

0.04

Instability Model B (S.E.)

β

5.04 (0.91) 0.01 (0.01)

0.36⁎⁎⁎ 0.09

− 0.01 (0.01) 0.09 (0.04)

− 0.04

0.02 (0.01)

0.24⁎⁎⁎

− 0.07 0.05

0.14⁎

4. Discussion Prior research demonstrated a link between foster care status and negative mental health outcomes (Courtney et al., 2001; Courtney &

⁎⁎⁎

p ≤ 0.001.

164

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

et al., 2001; Geenen & Powers, 2007; Stott, 2012; but see Berger et al., 2009). Our findings build on this work by showing that instability is a stronger associate of mental health problems, specifically PTSD, than foster care status itself. However, we must interpret these results with caution as it remains possible that children who enter care with lower mental health functioning are moved to different homes and schools more often than children with better mental health. Causation cannot be inferred from our study. Nevertheless, our results indicated that foster care status is associated with increased residential and school instability, which in turn predict self-reports of greater PTSD symptoms and somewhat greater depression/anxiety symptoms. However, in our analyses, the direct paths between foster care status and the mental health outcomes were not significant. These results support the possibility that some of the negative mental health outcomes commonly observed among foster youth may be at least in part a consequence of the lack of residential and educational stability experienced by this population, rather than a direct effect of foster care status broadly defined. Furthermore, consistent with previous research showing that children in foster care exhibit higher rates of PTSD compared to the general population (Keller, Salazar, & Courtney, 2010; Pecora et al., 2009), our results from the group comparisons indicated that instability predicted more severe PTSD symptoms in the foster youth but not in the at-risk comparison group with no history of foster care. These results may imply that the impact of instability is not solely quantitative but qualitative as well; instability for foster youth versus nonfoster youth increases PTSD symptomology likely as a result of the type of disruptions. When foster youth change homes and schools, the instability permeates throughout their lives, both at home with their caregivers, and at school with their peers; however, instability for nonfoster youth is more likely to be constrained to peer disruptions as instability is likely less often accompanied by a change in caregiver. Our study leaves open a number of important questions, such as whether instability and the associated loss is a trauma in and of itself, if instability compounds previous trauma, and if instability results in a lack of support/resources for children to cope with and resolve past trauma (e.g., of abuse and/or neglect). The observed increase in vulnerability to poor mental health associated with greater instability did not extend to depression/anxiety problems. Although this finding is in contrast to studies indicating that children in foster care are more likely to experience mental health problems compared to the general population (Pecora et al., 2009), given that our comparison group was comprised of at-risk youth rather than age-matched youth from the general population, our results may not be surprising. Furthermore, although the previous literature has been somewhat mixed, at least one prior study also reported rates of emotional functioning problems such as depression in foster care youth that were comparable to the lifetime prevalence of those among older adolescents in the general population (Keller et al., 2010). The lack of difference might also be the result of bias potentially reflected in selfreport measures of mental health. As placement into foster care may disrupt the ongoing bond between child and caregiver (Casanueva et al., 2014) and because early attachment may remain stable until adulthood (Waters et al., 2000), it was proposed that youth in foster care would have more anxious and/or more avoidant attachment than youth without foster care histories. Contrary to our prediction, neither foster care status nor instability predicted attachment insecurity. It is possible that the foster and nonfoster children's attachment securities and insecurities were based on early life experiences, and that for foster youth, these experiences largely predated their foster care placements. It is also possible that the foster care placements did not disrupt the attachment system, at least to the extent that foster parent sensitivity and commitment to being a foster caretaker helps to alleviate the children's distress related to being removed from the home (Zeanah, Berlin, & Boris, 2011). Finally, it is possible that the self-report nature of the ECR measure made

Table 4 Path estimates for groups. Foster care

Instability → attachment avoidancea Instability → attachment anxietya Instability → depression/anxiety Instability → PTSD symptomsb

Nonfoster care

B (S.E.)

β

B (S.E.)

β

0 (0)

0

0 (0)

0

0 (0) 0.09 (0.05) 0.03 (0.01)

0 0.15t

0 (0) − 0.05 (0.09) 0.00 (0.01)

0 − 0.06

0.28⁎⁎⁎

− 0.02

p ≤ 0.10. p ≤ 0.001. a Path estimates between instability and attachment were constrained to 0. b Path estimates between instability and PTSD symptoms are significantly different between the two groups. t

⁎⁎⁎

Dworsky, 2006; Havlicek et al., 2013). Based on such findings, our first hypothesis was that foster care (compared to nonfoster care) status would be associated with higher levels of mental health problems, including more symptoms of PTSD, depression/anxiety, and attachment insecurities. Our analyses, however, did not uncover such relations. Instead our findings indicated that instability, as measured by combining placement and school movements, is associated with increased PTSD symptoms for foster youth. These findings are consistent, in part, with our fourth hypothesis, as tested by our group comparisons, that greater instability as measured by a higher number of homes lived in and schools attended predicts higher levels of PTSD symptoms in foster youth, although the findings did not extend to nonfoster youth. The lack of support for our first hypothesis is possibly due to the similarities between our foster and at-risk nonfoster groups. Comparing foster youth to at-risk nonfoster youth from high-poverty at-risk neighborhoods rather than to nonfoster youth from middle-class neighborhoods may have eliminated the significant main effects typically seen in research. Our foster care sample, although at higher risk than our control group for difficulties such as homelessness and teen pregnancy (Hobbs, Bederian-Gardner, Ogle, & Goodman, 2017), selfreported relatively good mental health overall. Given that our foster youth sample may have been higher functioning than average foster youth, it is nonetheless striking that residential and school instability had predictive effects in this sample. Children typically enter foster care after severe negative experiences with their birth families. These experiences can result in both internalizing and externalizing behavior problems that threaten placement stability (Schofield & Beek, 2005). Schofield and Beek describe the joint effort needed to keep foster children in stable placements. For example, foster parents must understand that the children will likely have difficulty trusting them and forming a secure attachment to them, and that it will take time for the foster parents and children to bond. Integration into the foster home, as assessed by the children's ability to form positive relationships with the foster family members and vice versa, predicts decreased placement disruption, including among youth with behavioral problems (Leathers, 2006), whereas older age of children, externalizing behavior, separation from siblings, and starting placement in residential care are among the predictors of increased placement instability (Rock, Michelson, Thomson, & Day, 2013). Services such as respite care (Schofield & Beek, 2005), special high school “academies” (Lawler, Sayfan, Goodman, Narr, & Cordon, 2014), and training and interventions to assist with developing secure attachments in foster children (Dozier et al., 2006; Dozier et al., 2009) are designed in part to help decrease frequent placements. Previous research indicates that instability in the home, school, and neighborhood is detrimental to foster youth and that stability is essential for foster youths' positive outcomes (Blome, 1997; Courtney 165

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

attachment insecurities difficult to detect. Had we used a different measure of attachment, such as the Adult Attachment Interview (George, Kaplan, & Main, 1985), we might have found predicted results. However, as the goal was to measure current levels of attachment and not current views of prior attachment (Bartholomew & Shaver, 1998), the ECR was the preferred measure in this instance. Future research on foster care and attachment should consider type and timing of placements in addition to the number of moves as those factors may be predictive of later self-reported attachment insecurities.

Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. In J. A. Simpson, & W. S. Rholes (Eds.). Attachment theory and close relationships (pp. 46–76). New York, NY: Guilford Press. Browne, M. W., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods & Research, 21, 230–258. http://dx.doi.org/10.1177/ 0049124192021002005. Casanueva, C., Dozier, M., Tueller, S., Dolan, M., Smith, K., Webb, M. B., ... Harden, B. J. (2014). Caregiver instability and early life changes among infants reported to the child welfare system. Child Abuse & Neglect, 38, 498–509. http://dx.doi.org/10.1016/ j.chiabu.2013.07.016. Casey Family Programs (2007). Educating children in foster care: The McKinney-Vento and no child left behind acts. Chafee (2008). National Youth in transition database: Final rule. Federal Register, 73, 10338–10378. Chambers, R. M., Crutchfield, R. M., Willis, T. Y., Cuza, H. A., Otero, A., & Carmichael, H. (2017). Perspectives. Former foster youth refining the definition of placement moves. Children and Youth Services Review, 73, 392–397. http://dx.doi.org/10.1016/j. childyouth.2017.01.010. Child Welfare Information Gateway (2013). How the child welfare system works. Conradi, H. J., & de Jonge, P. (2009). Recurrent depression and the role of adult attachment: A prospective and a retrospective study. Journal of Affective Disorders, 116, 93–99. http://dx.doi.org/10.1016/j.jad.2008.10.027. Courtney, M. E., & Dworsky, A. (2006). Early outcomes for young adults transitioning from out-of-home care in the USA. Child & Family Social Work, 11, 209–219. http:// dx.doi.org/10.1111/j.1365-2206.2006.00433. Courtney, M. E., 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. Cunningham, M. J., & Diversi, M. (2012). Aging out: Youths' perspectives on foster care and the transition to independence. Qualitative Social Work, 12, 587–602. http://dx. doi.org/10.1177/1473325012445833. Dolan, M., Casanueva, C., Smith, K., & Ringeisen, H. (2013). NSCAW child well-being spotlight: More than one quarter of children placed out of home experience placement disruption in the first 18 months after a maltreatment investigation. OPRE Report #2013-05. Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. Dozier, M., Lindhiem, O., Lewis, E., Bick, J., Bernard, K., & Peloso, E. (2009). Effects of a foster parent training program on young children's attachment behaviors: Preliminary evidence from a randomized clinical trial. Child and Adolescent Social Work Journal, 26, 321–332. Dozier, M., Peloso, E., Lindhiem, O., Gordon, M. K., Manni, M., Sepulveda, S., ... Levine, S. (2006). Developing evidence-based interventions for foster children: An example of a randomized clinical trial with infants and toddlers. Journal of Social Issues, 62, 767–785. http://dx.doi.org/10.1111/j.1540-4560.2006.00486. Ellermann, C. R. (2007). Influences on the mental health of children placed in foster care. Family & Community Health, 30, S23–S32. http://dx.doi.org/10.1097/01.FCH. 0000264878.50569.a3. Farineau, H. M., Wojciak, A. S., & McWey, L. M. (2013). You matter to me: Important relationships and self-esteem of adolescents in foster care. Child & Family Social Work, 18, 129–138. http://dx.doi.org/10.1111/j.1365-2206.2011.00808.x. Fernandez, E. (2007). How children experience fostering outcomes: Participatory research with children. Child & Family Social Work, 12, 349–359. http://dx.doi.org/10. 1111/j.1365-2206.2006.00454.x. Festinger, T. (1983). No one ever asked us…A postscript to foster care. New York, NY: Columbia University Press. Fraley, R. C., Fazzari, D. A., Bonanno, G. A., & Dekel, S. (2006). Attachment and psychological adaptation in high exposure survivors of the September 11th attack on the World Trade Center. Personality and Social Psychology Bulletin, 32, 538–551. http:// dx.doi.org/10.1177/0146167205282741. Fraley, R. C., & Shaver, P. R. (1997). Adult attachment and the suppression of unwanted thoughts. Journal of Personality and Social Psychology, 73, 1080–1091. http://dx.doi. org/10.1037//0022-3514.73.5.1080. Fraley, R. C., & Shaver, P. R. (2000). Adult romantic attachment: Theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4, 132–154. http://dx.doi.org/10.1037//1089-2680.4.2.132. Geenen, S., & Powers, L. E. (2007). “Tomorrow is another problem” the experiences of youth in foster care during their transition into adulthood. Children and Youth Services Review, 29, 1085–1101. http://dx.doi.org/10.1016/j.childyouth.2007.04.008. George, C., Kaplan, N., & Main, M. (1985). Adult attachment interview protocol (Unpublished manuscript)Berkeley: University of California. Goldhaber-Fiebert, J. D., Bailey, S. L., Hurlburt, M. S., Zhang, J., Snowden, L. R., Wulczyn, F., ... Horwitz, S. M. (2012). Evaluating child welfare policies with decisionanalytic simulation models. Administration and Policy in Mental Health, 39, 466–477. http://dx.doi.org/10.1007/s10488-011-0370-z. Hango, D. W. (2006). The long-term effect of childhood residential mobility on educational attainment. The Sociological Quarterly, 47, 631–661. http://dx.doi.org/10. 1111/j.1533-8525.2006.00061.x. Havlicek, J. R., Garcia, A. R., & Smith, D. C. (2013). Mental health and substance use disorders among foster youth transitioning to adulthood: Past research and future directions. Children and Youth Services Review, 35, 194–203. http://dx.doi.org/10. 1016/j.childyouth.2012.10.003. Hazan, C., & Shaver, P. R. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52, 511–524. http://dx.doi.org/10. 1037/0022-3514.52.3.511. Hobbs, S. D., Bederian-Gardner, D., Ogle, C. M., & Goodman, G. S. (2017, April).

4.1. Limitations Data on mental health and attachment were collected for only a subsample of foster youth surveyed in Phase 1 of NYTD. These foster youth were willing to answer the NYTD required questions and furthermore to answer additional questions. This may have resulted in a self-selection bias of foster youth with better mental health and more secure attachments compared to youth who were unwilling to answer the additional questions. Even so, our results indicate that the youth in our study experienced some mental health problems. Of importance, given that our data are cross-sectional (not longitudinal), we cannot determine whether early PTSD and depression/anxiety symptomology and/or more insecure attachment triggered multiple moves. 5. Conclusions Our findings build on previous research on psychological health in foster care youth by showing that residential and school instability is a stronger associate than foster care status itself of mental health difficulties, specifically symptoms of PTSD. Not only are frequent placement disruptions problematic for foster children, they are also costly (Goldhaber-Fiebert et al., 2012). Thus these results arguably provide an added incentive to work to increase residential and school stability for foster youth in an effort to provide them with the consistency and quality of care they need and deserve. References Arbuckle, J. L., & Wothke, W. (1999). Amos users' guide, Version 4.0. Chicago, IL: Small Waters Corporation. Barber, J. G., & Delfabbro, P. H. (2005). Children's adjustment to long-term foster care. Children and Youth Services Review, 27, 329–340. http://dx.doi.org/10.1016/j. childyouth.2004.10.010. Bartholomew, K., & Horowotz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226–244. http://dx.doi.org/10.1037/0022-3514.61.2.226. Bartholomew, K., & Shaver, P. R. (1998). Methods of assessing adult attachment. Do they converge? In J. A. Simpson, & W. S. Rholes (Eds.). Attachment theory in close relationships (pp. 25–45). New York, NY: Guilford Press. Belsky, J., & Fearon, R. M. P. (2008). Precursors of attachment security. In J. Cassidy, & P. R. Shaver (Eds.). Handbook of attachment: Theory, research, and clinical applications (pp. 295–316). (2nd ed.). New York, NY: Guilford Press. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246. Berger, L. M., Bruch, S. K., Johnson, E., James, S., & Rubin, D. (2009). Estimating the “impact” of out-of-home placement on child well-being: Approaching the problem of selection bias. Child Development, 80, 1856–1876. http://dx.doi.org/10.1111/j.14678624.2009.01372.x. Bergin, C., & Bergin, D. (2009). Attachment in the classroom. Educational Psychology Review, 21, 141–170. http://dx.doi.org/10.1007/s10648-009-9104-0. 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, 41–53. http://dx.doi.org/10.1023/ A:1024592813809. Bowlby, J. (1969). Attachment and loss vol 1: Attachment. New York, NY: Basic Books. Bowlby, J. (1977). The making and breaking of affectional bonds. I. Aetiology and psychopathology in the light of attachment theory. An expanded version of the fiftieth Maudsley lecture, delivered before the Royal College of Psychiatrists, 19, November, 1976. The British Journal of Psychiatry, 130, 201–210. http://dx.doi.org/10.1192/bjp. 130.3.201. Bowlby, J. (1980). Attachment and loss vol. 3: Loss, sadness and depression. London: The Hogarth Press. Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human development. New York, NY: Basic Books.

166

Children and Youth Services Review 84 (2018) 159–167

D. Bederian-Gardner et al.

Trauma: Theory, Research, Practice, and Policy, 7, 324–332. http://dx.doi.org/10. 1037/tra0000015. Pecora, P. J., White, C. R., Jackson, L. J., & Wiggins, T. (2009). Mental health of current and former recipients of foster care: A review of recent studies in the USA. Child & Family Social Work, 14, 132–146. http://dx.doi.org/10.1111/j.1365-2206.2009. 00618.x. Reilly, T. (2003). Transition from care: Status and outcomes of youth who age out of foster care. Child Welfare, 82, 727–746. Rock, S., Michelson, D., Thomson, S., & Day, C. (2013). Understanding foster placement instability for looked after children: A systematic review and narrative synthesis of quantitative and qualitative evidence. British Journal of Social Work, 45, 177–203. http://dx.doi.org/10.1093/bjsw/bct084. Rubin, D., O'Reilly, A., Luan, X., & Localio, A. (2007). The impact of placement stability on behavioral well-being for children in foster care. Pediatrics, 119, 336–344. Salazar, A. M., Keller, T. E., Gowen, L. K., & Courtney, M. E. (2013). Trauma exposure and PTSD among older adolescents in foster care. Social Psychiatry and Psychiatric Epidemiology, 48, 485–551. http://dx.doi.org/10.1007/s00127-012-0563-0. Sameroff, A. J., & Seifer, R. (1990). Early contributors to developmental risk. In J. Rolf, A. S. Masten, D. Cicchetti, K. H. Neuchterlein, & S. Swintraub (Eds.). Risk and protective factors in the development of psychopathology (pp. 52–66). New York, NY: Cambridge University Press. Schofield, G., & Beek, M. (2005). Risk and resilience in long-term foster-care. The British Journal of Social Work, 35, 1283–1301. http://dx.doi.org/10.1093/bjsw/bch213. Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking and support giving within couples in an anxiety-provoking situation: The role of attachment styles. Journal of Personality and Social Psychology, 62, 434–446. http://dx.doi.org/10.1037/ 0022-3514.62.3.434. Solomon, Z., Ginzburg, K., Mikulincer, M., Neria, Y., & Ohry, U. (1998). Coping with war captivity: The role of attachment style. European Journal of Personality, 12, 271–285. http://dx.doi.org/10.1002/(SICI)1099-0984(199807/08)12:4<271::AIDPER309>3.0.CO;2-U. Stott, T. (2012). Placement instability and risky behaviors of youth aging out of foster care. Child and Adolescent Social Work Journal, 29, 61–83. http://dx.doi.org/10.1007/ s10560-011-0247-8. Thompson, R., & Zelazo, P. D. (2013). Attachment theory and research: Precis and prospect. The Oxford handbook of developmental psychology. Vol. 2. The Oxford handbook of developmental psychology (pp. 191–216). Oxford UK: Oxford University Press. Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maximum likelihood factor analysis. Psychometrika, 38, 1–10. http://dx.doi.org/10.1007/BF0229117. U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau (2016). Child maltreatment 2014. Available from http://www.acf.hhs.gov/programs/cb/researchdata-technology/statistics-research/child-maltreatment. Unrau, Y. A., Seita, J. R., & Putney, K. S. (2008). Former foster youth remember multiple placement moves: A journey of loss and hope. Children and Youth Services Review, 30, 1256–1266. http://dx.doi.org/10.1016/j.childyouth.2008.03.010. Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim, L. (2000). Attachment security in infancy and early adulthood: A twenty-year longitudinal study. Child Development, 684–689. http://dx.doi.org/10.1111/1467-8624.00176. Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007). The experiences in close relationship scale (ECR)-short form: Reliability, validity, and factor structure. Journal of Personality Assessment, 88, 187–204. http://dx.doi.org/10.1080/ 00223890701268041. Weinfield, N. S., Sroufe, L. A., & Egeland, B. (2000). Attachment from infancy to early adulthood in a high-risk sample: Continuity, discontinuity, and their correlates. Child Development, 71, 695–702. http://dx.doi.org/10.1111/1467-8624.00178. Wothke, W. (2000). Longitudinal and multi-group modeling with missing data. In T. D. Little, K. U. Schnabel, & J. Baumert (Eds.). Modeling longitudinal and multilevel data: Practical issues, applied approaches, and specific examples (pp. 219–240). Hillsdale, NJ: Lawrence Erlbaum Associates. Zeanah, C. H., Berlin, L. J., & Boris, N. W. (2011). Practitioner review: Clinical applications of attachment theory and research for infants and young children. Journal of Child Psychology and Psychiatry, 52, 819–833. http://dx.doi.org/10.1111/j.14697610.2011.02399.

Emancipating foster youth: Factors influencing outcomes and mental health. Paper presented at the annual meeting of the Western Psychological Association, Sacramento, CA. Howes, C., & Ritchie, S. (1999). Attachment organizations in children with difficult life circumstances. Development and Psychopathology, 11, 251–268. http://dx.doi.org/10. 1017/S0954579499002047. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55. http://dx.doi.org/10.1080/ 10705519909540118. Jelleyman, T., & Spencer, N. (2008). Residential mobility in childhood and health outcomes: A systematic review. Journal of Epidemiology & Community Health, 62, 584–592. http://dx.doi.org/10.1136/jech.2007.060103. Keller, T. E., Cusick, G. R., & Courtney, M. E. (2007). Approaching the transition to adulthood: Distinctive profiles of adolescents aging out of the child welfare system. The Social Service Review, 81, 453–484. http://dx.doi.org/10.1086/519536. Keller, T. E., Salazar, A. M., & Courtney, M. E. (2010). Prevalence and timing of diagnosable mental health, alcohol, and substance use problems among older adolescents in the child welfare system. Children and Youth Services Review, 32, 626–634. http:// dx.doi.org/10.1016/j.childyouth.2009.12.010. King, L. A., King, D. W., Leskin, G. A., & Foy, D. W. (1995). The Los Angeles symptom checklist: A self-report measure of posttraumatic stress disorder. Assessment, 2, 1–17. Kline, R. B. (2011). Principles and practice of structural equation modeling (3rd ed.). New York, NY: Guilford Press. Kumar, A., & Sharma, S. (1999). A metric measure for direct comparison of competing models in covariance structure analysis. Structural Equation Modeling: A Multidisciplinary Journal, 6, 169–197. http://dx.doi.org/10.1080/ 10705519909540127. Lawler, M. J., Sayfan, L., Goodman, G. S., Narr, R., & Cordon, I. M. (2014). Comprehensive residential education: A promising model for emerging adults in foster care. Children and Youth Services Review, 38, 10–19. Leathers, S. J. (2006). Placement disruption and negative placement outcomes among adolescents in long-term foster care: The role of behavior problems. Child Abuse & Neglect, 30, 307–324. http://dx.doi.org/10.1016/j.chiabu.2005.09.003. Leonard, S., Stiles, A. A., & Gudiño, O. (2016). School engagement of youth investigated by child welfare services: Associations with academic achievement and mental health. School Mental Health, 8, 386–398. http://dx.doi.org/10.1007/s12310-0169186-z. Lewis, M., Feiring, C., & Rosenthal, S. (2000). Attachment over time. Child Development, 71, 707–720. http://dx.doi.org/10.1111/1467-8624.00180. Marcynyszyn, L. A., Evans, G. W., & Eckenrode, J. (2008). Family instability in early and middle adolescence. Journal of Applied Developmental Psychology, 29, 380–392. Mikulincer, M., & Florian, V. (1998). The relationship between adult attachment styles and emotional and cognitive reactions to stressful events. In J. A. Simpson, & W. S. Rholes (Eds.). Attachment theory in close relationships (pp. 143–165). New York, NY: Guildford Press. Mikulincer, M., Florian, V., & Weller, A. (1993). Attachment styles, coping strategies, and post-traumatic psychological distress: The impact of the Gulf War in Israel. Journal of Personality and Social Psychology, 64, 817–826. http://dx.doi.org/10.1037/00223514.64.5.817. Mikulincer, M., & Shaver, P. R. (2016). Attachment in adulthood: Structure, dynamics, and change. New York, NY: Guilford Press. Mikulincer, M., Shaver, P. R., & Pereg, D. (2003). Attachment theory and affect regulation: The dynamics, development, and cognitive consequences of attachment-related strategies. Motivation and Emotion, 27, 77–102. http://dx.doi.org/10.1023/ A:1024515519160. Mikulincer, M., Shaver, P. R., & Solomon, Z. (2015). An attachment perspective on traumatic and posttraumatic reactions. In M. P. Safir, H. S. Wallach, & A. Rizzo (Eds.). Future directions in post-traumatic stress disorder: Prevention diagnosis and treatment (pp. 79–96). New York, NY: Springer. 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 & Neglect, 24, 1363–1374. http://dx.doi.org/10.1016/ S0145-2134(00)00189-7. Ogle, C. M., Rubin, D. C., & Siegler, I. C. (2015). The relation between insecure attachment and posttraumatic stress: Early life versus adulthood traumas. Psychological

167