The Role of Cultural Dissimilarity Factors on Child Adjustment Following Foster Placement

Abstract

Understanding the cultural factors associated with children's experiences in foster care is important because they may contribute to child psychological adjustment to foster placement. Despite considerable public policy debate of the role of ethnicity on foster placement decisions, there are virtually no empirical studies about the contribution of cultural dissimilarity factors on child psychological adjustment such as internalizing and externalizing problems shortly after children enter non-kinship placement.

Using a sample of N =106 ethnic minority children (clustered in 62 families), we hypothesized that the number (ranging from 0 to 5) and types (i.e., ethnic status, country of birth, and spoken language) of cultural dissimilarity factors between biological and foster families contribute to child internalizing symptoms (CDI depression and LSD loneliness) and externalizing problems (ECBI conduct) after considering family and agency clustering and adjusting for confound variables (child age, gender, and severity of child maltreatment).

Results showed that a higher number of dissimilar types and certain types contributed to lower scores in child psychological adjustment. Dissimilar ethnic status between caregivers contributes to CDI depression and LSD loneliness symptoms while dissimilar spoken language between caregivers contributed to ECBI conduct problems in the foster home. These results inform the public policy debate of transethnic placements for children involved in the foster care system.

In 2008, nearly half a million (408,425) children were placed in foster care (Adoption and Foster Care Analysis Report [AFCARS], 2010). While placing a child in foster care provides a temporary safe environment for abused or neglected children, all too often, their life is tumultuous and unstable. Shortly after children are removed from their homes and over the following three years about half (46%) meet diagnosis for separation anxiety disorder (Linares et al., 2010) suggesting that children experience considerable distress related to family loss and separation. Children placed in foster homes are at a greater risk of delayed development, have problems with behavioral regulation, and suffer from higher rates of internalizing and disruptive behavior disorders (Linares et al., 2010; Chipungu & Bent-Goodley, 2004; Pecora et al., 2000; Russell, Cutrona, Rose, & Yurko, 1984). After placement in foster homes, many children continue to experience social isolation and family instability (Bass, Shields & Behrman, 2004; Pecora et al., 2000).

Responding to historical gaps in the adoption of African-American children from foster homes, the Multiethnic Placement Act amended by the Interethnic Adoption Provision of 1996 (MEPA-IEP; http://www.acf.hhs.gov/programs/cb/pubs/mepa94/mepachp3.htm) was designed to prohibit routine consideration of race, color, and national origin in the foster placement process as a ways to improve adoption rates of ethnically minority children freed for adoption. As a remedy to high rates of minority children lingering in the foster care system, this legislation mandated that cultural factors may only be considered on an individualized basis when necessary to protect the best interests of the child. MEPA-IEP also encourages the active and diligent recruitment of foster and adoptive parents of all backgrounds that reflect the ethnic and racial diversity of children for whom foster and adoptive homes are needed. Following MEPA-IEP legislation, our local child welfare system endorsed the notion that `a placement must be reflective of, and responsive to, a child's specific culture, religion, and background (Principle 6c; http://www.nyc.gov/html/acs/downloads/pdf/stats_placement_report.pdf; retrieved May 20, 2011); consistent with this principle in 1999, 76% of children in NYC were placed in ethnically `compatible homes' (http://www.nyc.gov/html/acs/downloads/pdf/stats_placement_report.pdf; retrieved May 20, 2011).

Ethnic compatibility of biological and foster families may facilitate the child's positive ethnic identity. Research suggests that individuals identify themselves ethnically on the basis of appearance and group membership (Phinney, 2003; Phinney, 2007). Ethnic identity is “that part of an individual's self-concept which derives from [his] knowledge of [his] membership of a social group (or groups) together with the value and emotional significance attached to that membership” (Phinney, 2007). Ethnic identity is an aspect of self-concept which begins to develop in early childhood and serves as a protective factor enabling individuals to be resilient against social adversity and discrimination (Phinney 2003). A stronger and more positive connection to an individual's ethnic group predict higher self-esteem, academic achievement, psychological adjustment, coping abilities, and lower levels of depression and loneliness (White et al., 2008; Umana-Taylor, Bhanot, & Shin, 2006; Hollingsworth, 1997).

Ethnic dissimilarity in foster placement may also contribute to fewer ethnically compatible social networks for the child (Peplau & Perlman, 1982). Diminished social support networks may leave children unprotected to buffer symptoms of loneliness, social isolation, and depression (Sumer, Poyrazli, & Grahame, 2008). Children in foster care are often conflicted by dual loyalties as they wish to develop a sense of belongingness which may further interfere with positive ethnic identity formation (White et. al, 2008). For ethnic minorities, the development of a positive ethnic identity is more complex because beyond navigating new interpersonal relationships, individuals must also deal with their minority status within the majority culture (White et. al., 2008).

Cultural dissimilarity, particularly in the use of spoken language, may also be a risk factor for increased child conduct problems because cooperative parenting practices between biological and foster caregivers may be more challenging when caregivers do not share the same ethnic or language background. Unknown to each other until the placement event, the visiting biological parent and the foster caregiver must learn to share the child's affection and loyalties, acknowledge their parental differences, and negotiate disagreements involving permanency goals, family visitation, and daily routines such as discipline, homework, grooming, and recreation activities. Caregivers who share a similar ethnic background, country, and language may negotiate these co-parenting tasks easier, while dissimilar parents (on the basis of ethnicity, country of origin, or spoken language) may find these new tasks more difficult. Positive co-parenting practices (e.g., cooperation, negotiated conflict, and low triangulation) has been found to be related with fewer conduct problems (Montalto et al., 2011; Linares et al., 2010; Linares et al., 2006).

Despite the critical need to inform the public policy debate in the child welfare system, controlled studies of the contribution of cultural dissimilarity factors (types) on child adjustment to foster care remain rare. This research is a step in this direction. In this study, cultural dissimilarity types in foster care refers to the mismatch or incompatibility regarding ethnic self-identification, country of birth, and preferred language spoken between the child and foster parent, and the biological and foster parent. The purpose of this study is to investigate the contribution of cultural dissimilarity between biological and foster families in types (i.e., ethnic background, country of birth, and spoken language) and number (i.e., the number of different types) on reported child emotional and conduct problems in the foster home. We hypothesize that cultural dissimilarity factors (separately and in combination) between biological and foster families will be associated with child depression symptoms, feelings of loneliness, and conduct problems shortly after entering placement. We adjust for clustering in family (most children in the study were siblings), and agency (children were drawn from foster care sites) related to sampling procedures, demographic factors (age, gender), and the severity of child maltreatment history. The study advances existing developmental science by examining the association between cultural dissimilarities (in types and number of different types) between families and child behavior at a time of initial family fragmentation (i.e., foster placement) in a unique population of high-risk children.

Methods

Participants

The sample was selected from participants in a longitudinal study of N = 252 children (clustered in 95 families) which examined foster placement shifts, sibling relationship quality, and child outcomes in foster care (Linares et al., 2007). The children in the larger study entered foster care at one of thirteen participating foster care agencies in NYC; they were drawn from 560 sibling groups consecutively recruited during a 3-year recruitment period. Reflecting nationwide trends, in NYC over half of children (57.3%) enter care as a part of a sibling group (http://www.nyc.gov/html/acs/downloads/pdf/outcomes/ind2_quickview.pdf). From identified sibling groups, 20% were eligible for enrollment (n=95); 2% refused.

The study sample is comprised of N = 106 children (clustered in 62 families: 28 with one child, 24 with two children; and 10 with three children) drawn from 9 foster care agencies. Children were between ages 7–15 and were selected because they were old enough to provide self-reports of internalizing symptoms (CDI depression and LSD loneliness) at the time of the baseline visit. On average, baseline assessments were obtained within 3-months of initial placement. Children in the study who were siblings were placed together in the same non kinship (i.e., with a caregiver who was a non-relative to the child) foster home.

Children mean age was 10.47 years (SD =1.95) ranging from 7.7 to 15.0 years. The gender distribution of this sample (55% male) resembles the national foster care population (53% male). According to official records which conformed to the NYS legal definitions of child maltreatment (neglect, physical abuse, and sexual abuse), children were classified as 76% neglected, 20%, physically abused, and 4% sexually abused. Children came from family households of low socioeconomic status. Foster parent had an average of < 12 years of education; approximately one third were on public assistance; and approximately one third resided in government-sponsored housing.

Procedures

Written informed consent was gathered from biological parent, foster parent and child assent was obtained from children ≥ 7-years of age according to approved IRB protocols from New York University, and local and state Department of Social Services (foster care agency approval was also granted for those agencies with IRB bodies). Multi-informant data were gathered via face-to-face interviews in the foster home from the foster parent, the child, and the biological parent (interviewed in her home or agency). Study outcomes included self-report measures of child depressive symptoms and peer loneliness; and foster parent-reported conduct problems in the foster home. Type of child maltreatment was gathered from Child Protective System (CPS) official records kept at the agency.

Measures

Predictor variables

Cultural dissimilarity between biological and foster families was assessed by type (mismatched ethnicity, country of birth, and spoken language) and number of dissimilar types:

Ethnicity

Participants were asked to report on their ethnicity using one of the following categories: African-American, Latino, Caucasian, West-Indies, African, Asian, or Mixed. About one quarter of children (26%) identified themselves as mixed ethnicity. Among them, 46% were Latino and African-American, 33% were Latino or African-American and other, while 21% were Caucasian and other. Child ethnic self-identification data were obtained from the child verified by his/her biological parent while ethnic self-identification for the biological parent and foster parent was gathered from self-report. Ethnic Child-FP and ethnic FP-BioP mismatch were assigned each a category =1 while ethnic Child-FP and ethnic FP-BioP match were assigned a category =0. For children classified as of mixed ethnicity, we coded an ethnic child-FP mismatch only in cases where the FP ethnicity was different from any of the child's reported ethnicities. Pairs who reported being of mixed ethnicities were coded as = 0 (match).

Country of birth

Country of birth FP-BioP mismatch was assigned category = 1 if country of birth between caregivers differ and category = 0 if country of birth between caregivers was the same. Country of birth Child-FP mismatch was not assessed because all children were U.S. born.

Spoken language was defined as the language (English or Spanish) preferred at the time of the baseline interview. Bilingual caregivers who did not express a preference were interviewed in English. Language Child-FP mismatch between the child and the foster parent was assigned category =1; and language FP-BioP mismatch between the foster parent and the biological parent was assigned category =1. Language matches were assigned a category = 0.

Number of dissimilar types

Across the five dissimilarity types (2 ethnicity, 1 country of birth, 2 spoken language) scores ranged from 0 = zero dissimilar types to 5 =dissimilar for all types). No family had 5 dissimilar types. As an illustration, an English speaking Latino child has a US born biological mother who speak English; the foster parent is Latino, born in the Dominican Republic, and speaks Spanish; this child will receive a score of 3.

Covariates. Age, gender, and severity of child maltreatment were the study covariates. The Maltreatment Classification System (MCS, Barnett & Manly, 1993) measures the type, severity, and perpetrator of child maltreatment from information independently coded from official case record narratives. Narratives are coded under one or more of eight child maltreatment types (physical abuse 20%, sexual abuse 4%, neglect-failure to provide 28%, neglect-lack of supervision 46%, neglect-exposure to intimate partner violence 33%, emotional 12%, moral/legal 9%, and educational 16%) and rated by severity (on 1- to 5- point scale or undefined) and perpetrator. For this study we aggregated severity ratings across the eight types of maltreatment to create a continuous severity index of child maltreatment. Kappa coefficients were gathered for 10% of the sample yielding a mean Kappa coefficient of .850 across types of maltreatment.

Child outcomes

Depression symptoms

The Children's Depression Inventory (CDI, Kovacs, 1992) is a child-reported 27 item scale assessing feelings of loneliness, sadness, poor appetite and sleep, and thoughts about suicide during the past two weeks. The CDI total used here is a continuous measure which yields a total score. Alpha was .821.

Loneliness

The Loneliness and Social Dissatisfaction (Parker & Asher, 1993) Scale is modified version of a 24-item 5-point self-report questionnaire containing 16 items which focused on feelings of loneliness and social dissatisfaction at school and 8 filler items. Extra items were added to include the foster home. Items assess children feelings of loneliness, appraisal of current peer relationships, degree to which important relationship needs are being met, and their own social competence. In past research, loneliness total score ranged between 16 and 74 with a mean loneliness score of 29.44 (SD = 9.32). Alpha was .800.

Conduct problems

The Eyberg Child Behavior Inventory (ECBI; Eyberg & Pincus, 1999) is a parent reported measure for ages 2–17 used to assess conduct problems. The ECBI Total measure consists of 36 items which yield an intensity problem score; it has been shown to correlate with independent observations of children's behavior and to differentiate clinic-referred and non-clinical populations (Eyberg & Pincus, 1999). The measure is gathered from the foster parent and refers to behavior in the foster home. Alpha was .944.

Data Analyses

Descriptive analyses show the frequency distributions for predictor variables and means and SD for child outcome variables. To examine the contribution of cultural dissimilarity factors on child's initial psychological adjustment (CDI depression symptoms, LSD loneliness, and ECBI conduct problems), we conducted two main analyses using Generalized Linear Modeling (GENLIN) procedures in SPSS (version 17) that adjusts for family and agency clustering given that most children in the study were siblings and were drawn from different foster care sites. In the first regression analyses, we used number of cultural dissimilary types (0–5) as the predictor variable to examine the contribution of number of dissimilar types across five domains involving families' ethnicity, country of birth, and language on level of CDI depression, LSD loneliness, and ECBI conduct problems. In the second regression analyses, we used types of cultural dissimilarity as the predictor variables: ethnicity Child-FP, ethnicity FP-BioP, country of birth FP-BioP, language Child-FP, and language FP-BioP to evaluate the independent contribution to study outcomes when these variables were entered together in the models. In both analyses, covariates in the models included: child age, gender, and severity of maltreatment because they are known factors to influence levels of study outcomes (CDI depression, LSD loneliness, and ECBI conduct problems).

Results

Descriptive Analyses

Five ethnic groups were represented in our sample: African-American, Latino, Mixed (primarily of African-American or Latino ethnicities and Other), Caucasian, and West Indies. All children were born in the US, 73 biological mothers and 64 foster parents were US born. As seen in Table 1, 40% and 32% of foster parent and biological parent respectively were born in 16 different countries. Children were fluent in English and were interviewed in English. Based on language preference, 82% foster parents and 89% biological parents were interviewed in English. Types of cultural dissimilarity were positively intercorrelated for ethnicity and country of birth variables ranging from r = .69 to r = .30; while language Child-FP was negatively correlated to language FP-BioP (r = −.32; p <.001).

Table 1.

Cultural Composition of Study Families

Ethnicity	N (%)
	Child	Foster Parent	Biological Parent
African-American	42 (39.6)	61 (57.5)	54 (50.9)
Latino	30 (28.3)	26 (24.5)	38 (35.8)
Mixed	28 (26.4)	6 (5.7)	1 (0.9)
Caucasian	4 (3.8)	0 (0)	6 (5.7)
West-Indies	2 (1.9)	8 (7.5)	2 (1.9)
Country of birth
USA	106 (100)	64 (60.4)	73 (68.9)
PR		8 (7.5)	7 (6.6)
DR		12 (11.3)	6 (5.7)
Cuba			1 (0.9)
Ecuador		2 (1.9)
Honduras			1 (0.9)
Panama		3 (2.8)	2 (1.9)
Virgin Islands			2 (1.9)
Antigua		2 (1.9)
Bahamas			3 (2.8)
Barbados		1 (0.9)	3 (2.8)
Trinidad		1 (0.9)
England			2 (1.9)
Haiti		3 (2.8)	5 (4.7)
Jamaica		9 (8.5)	2 (1.9)
Guinea		1 (0.9)
Spoken language
English	106 (100)	87 (82.1)	94 (88.7)
Spanish		19 (17.9)	12 (11.3)

As seen in Table 2, 26% of families received a dissimilarity score of 0 (not dissimilar in any of the five types), while 7% received a score of 4. No families received a dissimilarity score of 5. On the average, the number of dissimilar types was M = 1.62 (SD =1.28). Cultural dissimilarity ranged from 17% for language Child-FP to 51% for country of birth FP-BioP. CDI total depression scores ranged from 0 to 31 with 16% of the sample scoring at the clinical range (T score ≥ 60) for CDI T total. LSD loneliness scores ranged from 16 to 79; and ECBI total conduct problems scores ranged from 39 to 224 with 43% of the children scoring at the clinical range (T score ≥ 60) for ECBI T Total.

Table 2.

Number of Culturally Dissimilar Types

Number	%
Zero	26
One	25
Two	18
Three	24
Four	7

GENLIN Analyses

In regard to number of dissimilar types, considering data clustering and above and beyond demographics (child age and gender) and the severity of maltreatment, number of dissimilar types predicted CDI depression (Wald Chi Square = 16.92, df 4, p <.01); LSD loneliness (Wald Chi Square = 12.58, df 4, p <.05); and ECBI conduct problems (Wald Chi Square = 10.20, df 4, p <.05).

In regard to specific types, considering data clustering and above and beyond demographics (child age and gender) and the severity of maltreatment, dissimilar ethnicity FP-BioP predicted CDI depression (B = −5.50.27; SE = 2.09, p =<.01); dissimilar ethnicity FP-BioP predicted LSD loneliness (B = −10.38; SE = 4.92, p < .05); and dissimilar language FP-BioP predicted ECBI conduct problems (B =−32.17, SE =16.15, p < .05).

Discussion

Although there is considerable debate about the merits of the MEPA-IEP public law (e.g., Donaldson Institute, 2008) to improve adoption rates for African-American children in foster care, it is largely unknown whether cultural compatibility factors in families involved in the foster care system contribute to child psychological well-being in foster placements. The aim of this study was to investigate the contribution of cultural dissimilarity variables on initial psychological adjustment to non-kinship foster homes among children ages 7–15. We hypothesized that cultural dissimilarity factors between biological and foster families along ethnic, country of birth, and spoken language types would be associated with higher levels of child depressive symptoms, feelings of loneliness and social dissatisfaction, and conduct problems in the foster home shortly after children are placed in non-kinship foster homes. Unlike studies of majority vs. minority children, our sample was predominantly composed of ethnic minority families of African-American and Latino backgrounds. The sample was multicultural, ethnically diverse, and heterogeneous within ethnic minority status, country birth, and language usage.

The hypotheses were supported for the impact of number of dissimilar types: the more the number of dissimilar cultural characteristics between families the more reported child problems; and partially supported for the impact of specific cultural dissimilarity factors after adjusting for family (siblings) and (agency) sampling procedures, and controlling for demographic factors (child age, gender), and severity of child maltreatment. Results showed that cultural mismatch between foster children and their caregivers has measurable negative effects on self-reports of child internalizing symptoms and foster parent reported level of child externalizing problems. The study suggests that ethnic dissimilarity largely contribute to child symptoms (i.e., internalizing) experienced following initial placement. For depressive symptoms and loneliness and social dissatisfaction outcomes, dissimilar ethnicity between biological and foster parent surfaces as detrimental perhaps because they create in the displaced child a lack of belongingness which reduced the child's ethnic identity and social connectedness to the foster home. Previous research has demonstrated that self-esteem buffers the negative effects of stress on depression, and that lower self-esteem was associated with more depressive symptoms (Hollingsworth, 1997).

For conduct problems, not surprisingly, a common language between biological and foster families is important because it is likely to facilitate direct communication between caregivers about the child's likes and dislikes, family routines and rules, and shared expectations all of which promotes behavioral adaptation to the foster home. Collaborative co-parenting is related to fewer conduct problems in the foster home (Montalto et al., 2011; Linares et. al., 2010).

Given that the sample was primarily ethnic minority, we were intrigued to find child outcome effects within this restricted range. Our findings are consistent with child welfare policies embraced by Child Welfare League of America, and the National Association of Black Social Workers that discourage transethnic adoptions of minority children into majority families for the case of African-American children and white adoptive parents; our findings extends support of this policy to the case of African-American and Latino children placed in minority although dissimilar foster homes.

Study limitations are in order. The sample size is small; the ethnic minority African- American/Latino subjects in this study reflect the unique multicultural diversity of NYC and may not generalize to other more homogeneous geographic regions. Race and ethnicity variables in the study were confounded. The study is crossectional of a single point shortly after entering in the foster home; it is possible that findings reflecting the detrimental role of ethnic mismatch diminish after the initial adjustment because ethnic similarity with an unfamiliar foster parent may be more important during times of high family instability. Other parental characteristics not considered here such as foster parental warmth and low hostility have an important role on child's behavior over time (Linares et al., 2010). The number of foster home changes and length of time in foster placement measured also affect the psychological well-being of children in the long run (Fisher, Kim & Pears, 2009). Future studies should include these and other potential influences that may account for the variation in child outcomes other than cultural dissimilar factors considered here. Future studies may also include direct measurements of ethnic self-identity, psychological resilience, the ethnic composition of other family members in the foster home, and religious practices. Finally, in future studies, data on transethnic placement should be collected over time, allowing for a more comprehensive representation on child outcomes.

The outcomes of this study, if duplicated, and more rigorously tested, have the potential to shape future foster placement decision legislation in order to make ethnic matching between foster child and caregiver one of the priority considerations. This study constitutes a step in the right direction. The use of standard assessments of emotional and behavioral functioning in a well-defined sample of children and multi-informant assessments collected from different sources reduce measurement error and are an important research step in informing this public debate in the foster care system. Findings of this study are intriguing and provide initial evidence of the promise of our multi-domain cultural construct pointing to similarities in ethnic self-identification (for internalizing symptoms) and language spoken in the foster home (for externalizing problems) as important domains that promote child psychological well-being. At the very least, these findings are important to promote dialogue and further research in this public policy debate.

Highlights.

Results showed that a higher number of dissimilar types and certain types contributed to lower scores in child psychological adjustment. Dissimilar ethnic status between caregivers contributes to CDI depression and LSD loneliness symptoms while dissimilar spoken language between caregivers contributed to ECBI conduct problems in the foster home. These results inform the public policy debate of transethnic placements for children involved in the foster care system.

Table 3.

Study Variables
Predictor variables
Dissimilar ethnicity	N(%)
Child-foster parent	35 (37)
Foster parent-BioP	39 (41)
Dissimilar country of birth
Foster parent-BioP	51 (51)
Dissimilar spoken language
Child-foster parent	17 (17)
Foster parent-BioP	23 (23)

Outcome variables	M (SD)	Range
CDI Total	10.16 (7.46)	0–31
LSD Loneliness	29.10 (12.51)	16–79
ECBI Conduct Problems	129.68 (46.36)	39–224

Note. BioP= Biological parent; CDI = Children's Depression Inventory; LSD =Loneliness and Social Dissatisfaction Scale. ECBI = Eyberg Children Behavior Inventory.

TABLE 4.

GENLIN Regressions Predicting Child Outcomes Based on Types of Cultural Dissimilarity

	CDI Depression				Child outcomes LSD Loneliness				ECBI Conduct problems
Predictor variables	B	SE	χ 2	P	B	SE	χ 2	P	B	SE	χ 2	P
Intercept	23.3	4.86	17.85	.000	39.00	9.56	16.66	.000	176.29	28.20	39.09	.000
Dissimilar ethnicity
Foster parent-BioP	−5.50	2.09	6.88	.01	−10.38	4.92	4.44	.05
Dissimilar spoken language
Foster parent-BioP									−32.17	16.15	3.97	.04

Note. Adjusted for data clustering, child age, gender, and maltreatment severity. BioP= Biological parent; FP = Foster parent; CDI = Children's Depression Inventory; LSD =Loneliness and Social Dissatisfaction Scale. ECBI = Eyberg Children Behavior Inventory.