Academic Functioning of Youth in Foster Care: The Influence of Unique Sources of Social Support

Abstract

Youth in foster care often experience more difficulty in school compared to their non-foster care peers. Difficulties exist across domains of academic functioning, including both performance (e.g., low grades) and behavioral health (e.g., high externalizing concerns) in school. One factor that has shown to be associated with positive academic functioning in the general population but remains to be comprehensively examined among youth in foster care is social support. This includes examining specific sources of support for youth in foster care and taking into consideration the context of the frequent placement disruptions many children in foster care experience. This study sought to determine which sources of social support are associated with academic functioning for youth in foster care by examining child-report of social support from parents, teachers, friends, and classmates in relation to school grades and teacher-reported behavioral health outcomes. Information on each source of social support was obtained from the self-report of 257 youth in foster care, and information on placement characteristics were obtained from child welfare casefiles. Teachers provided information on youth’s behavioral health in school, and academic grades were obtained from school records. Results suggested that youth reported teacher social support, as compared to parent, friend, or classmate social support, was most influential for both performance and behavioral health in school. Findings highlight the need for additional research on the important role of teachers for promoting academic success amongst youth in foster care, as well as the importance of placement changes in relation to academic functioning.

The academic functioning of youth in foster care is important, and the foster care system in which almost half a million youth in the US reside each year creates unique challenges for their academic development (U.S. Department of Health and Human Services, 2017). With regard to performance in school, research shows that youth in foster care are more likely to demonstrate failing grades and low standardized test scores than their peers who are not in foster care (Luke & O’Higgins, 2018; Trout et al., 2008). More than half of youth in foster care report failing a grade at least once, and nearly one-third repeat a grade at some point during their academic careers. Further, a high percentage, ranging from 30% to 50%, of youth in foster care are eligible for and placed in special education curriculum programs (Pecora, 2012). It is estimated that half of all youth in foster care are vulnerable to school failure and/or school dropout (Reilly, 2003; Zetlin et al., 2012). Additionally, youth in foster care display high rates of aggressive and noncompliant behaviors, as well as internalizing concerns, at school in comparison to non-foster peers (Gabrielli et al., 2015; Trout et al., 2008). These kinds of negative emotional and behavioral concerns in the classroom can further contribute to class failure and other school conduct problems, such as suspensions and bullying (McMillen et al., 2005).

According to Bronfenbrenner’s ecological systems theory (Bronfenbrenner, 1992), youth development is influenced by individual characteristics (e.g., genetics), as well as environmental (e.g., family and social relationships, community programs, placement changes) and systemic (e.g., state and federal laws) factors that are associated with academic functioning, including both performance and behavioral health (which for the purposes of this paper encompasses social, emotional, and behavioral functioning) in school. Further, the magnitude of problems in school among youth in foster care seems to vary considerably, depending on these individual and environmental factors (O’Higgins et al., 2017). Given that academic functioning can substantially influence later life outcomes (e.g., mental health, financial stability, and employment), identification and amelioration of poor academic functioning across domains should be a primary target for prevention and intervention efforts. One way to work toward this goal is through identification of modifiable environmental factors that positively support youths’ academic trajectories. While research to date has identified factors across various systems, one such environmental factor that has received minimal attention but that may be modifiable for youth in foster care is perceived social support. The current study examined youth perceived social support across multiple sources (e.g., teachers, friends) to determine which sources of social support may be most influential for both performance and behavioral functioning in school amongst youth in foster care.

Social Support

While definitions may vary, social support is often defined as the assistance given by other individuals to help the recipient manage or deal with a situation (Chu et al., 2010). This can include the provision of physical, psychological, or informational resources, or even the perception that these resources are available. Social support is a multidimensional construct, with perceived social support varying depending upon who is providing the support (i.e., source) or how they are providing it (i.e., type; Chu et al., 2010; Dunst & Trivette, 1988). Successful functioning across various domains of life often relies on the support of other close individuals; thus understanding the characteristics of support in an individual’s social network is essential for understanding an individual’s wellbeing. The empirical literature has demonstrated a clear link between social support and youth wellbeing, both in direct or main effects models, as well as through an indirect or buffering effects models. According to the main effects model of social support (the focus of the current study), social support has a direct influence on the wellbeing of a child, no matter a child’s history of exposure to stress, adversity, or other life experiences (Benhorin & McMahon, 2008; Cohen & Wills, 1985). Social support can also serve as a protective factor indirectly to promote resilience or wellbeing in youth who experience stress or adversity (Chu et al., 2010). In general, youth who report more social support tend to demonstrate better functioning across a number of areas, such as emotional, behavioral, social, and physical health (Chu et al., 2010; Heerde & Hemphill, 2018; Rueger et al., 2016).

Social support has also been associated with academic functioning, including both performance (e.g., grades, test scores) and behavioral health in the school setting (e.g., disciplinary problems, emotion regulation). As with other primary areas of development, functioning in the realm of academics is considerably influenced by youths’ social environment and the support they receive from other individuals (Lipschitz-Elhawi & Itzhaky, 2005; Welsh et al., 2001). For at-risk youth or youth who have experienced adversity, social support from close individuals may help promote motivation to engage in academics or to exhibit fewer problematic behaviors (e.g., skipping school), positive perceptions or satisfaction with school, and better performance in the classroom (Pan et al., 2019; Sterrett et al., 2011; Welsh et al., 2001). Moreover, findings on the benefits of social support appear to be fairly robust in relation to academic performance and behavioral health, as a positive relation between social support and academic functioning has been demonstrated across developmental stage, gender, socioeconomic status, and race/ethnicity (e.g., Benhorin & McMahon, 2008; DeGarmo & Martinez, 2006; Malecki & Demaray, 2006).

Although most research has examined social support overall (i.e., reporting on a single level or score of perceived social support from all other individuals), the specific source of social support can vary. This can include a range of individuals, such as individuals in the domains of home life (e.g., parents, siblings, extended family), school life (e.g., classmates, teachers, school counselors), social life (e.g., friends), and the community (e.g., social workers, neighbors, coaches), with the influence of some individuals extending across domains (Harter, 1985; Jackson & Warren, 2000). Evidence in general populations suggests that the specific sources of social support may be important in relation to academic functioning. For example, studies comparing teacher social support to other forms of social support (e.g., parent or classmate support) suggest that teacher social support may be uniquely important to how children perform and behave in school, as teacher social support may increase interest, engagement, and mastery in academics and curricular activities. Relatedly, Pan et al. (2017) found that teacher support demonstrated a stronger association with academic engagement as compared to parent social support, in a population of high-risk youth who had left school at some point or were close to being removed from school. Further, the positive association between social support and academic engagement was found to only remain for teachers when taking into account the child’s adverse life experiences (Pan et al., 2017).

However, evidence also suggests that both in-school (e.g., teacher, classmates/peers) and out-of-school (e.g., family members, friends) social support can positively influence how well youth do on tests or homework assignments and additionally influence their behavior in the classroom (e.g., DeGarmo & Martinez, 2006; Rueger et al., 2010). Parents and classmates or peers may provide emotional support that promotes general wellbeing in school, whereas teachers may provide informational support to promote academic performance (Hombrados-Mendieta et al., 2012). Taken together, previous research on academic functioning (including both performance and behavioral health in school) has shown that social support may be a key factor in determining how well youth function at school. This research has also demonstrated that, at least for youth in the general population, the individual providing the support may also be important to consider when examining the relation between academic functioning and social support, especially teacher social support.

Social Support and Youth in Foster Care

Youth in foster care represent a unique population of interest when considering the potential positive impact of perceived social support. Youth in foster care may experience, on average, less social support than their peers in the general population but perhaps need it more than their non-foster care peers. For example, youth in care often demonstrate difficulties forming intimate relationships with other adults and peers and generally have less family social support, in comparison to their non-foster care peers (Newton et al., 2000; Perry, 2006). Social support and positive relationships across all levels of a child’s social ecology, however, have been shown to not only buffer the effects of exposure to stress or adversity, but to be a powerful therapeutic agent following trauma exposure (Evans et al., 2013; Hambrick et al., 2019; Luby et al., 2019). Suboptimal social support may be an important reason why youth in care struggle with academic functioning and school connectedness. Thus, a first step toward improving the support received by children in care is to identify which sources of support are most useful in terms of promoting adaptive functioning in academic environments.

Although there has been some research on social support and academic functioning on youth in foster care, very little is known about how sources of social support contribute to academic functioning while youth are still in care (Stone, 2007). The majority of research examining the importance of social support among youth in foster care has focused on youth that have aged out or are transitioning out of foster care (Blakeslee, 2015; Curry & Abrams, 2015; Jones, 2014). Very few studies have focused on the role social support plays in the wellbeing of youth while still in care, such as with populations of school-age or younger adolescent foster care youth (Perry, 2006; Salazar et al., 2011). Research on youth who are currently in care may better elucidate how these concerns initially developed and are maintained, and thus provide insight into interventions to promote academic functioning. What research is available among youth still in care suggests that there is a positive association between social support and success in school functioning (Rosenfeld & Richman, 2003; Rosenfeld et al., 2000). That is, youth in foster care who have high levels of social support tend to demonstrate higher grades or test scores and report more positive behavioral health in school (e.g., academic motivation, school attendance, and fewer classroom behavior problems), as compared to those who report low levels of social support (e.g., Hyde & Kammerer, 2009; Rosenfeld & Richman, 2003).

Furthermore, evidence exists to suggest that the sources of social support youth in foster care receive matters. For example, Rosenfeld and Richman (2003) found that youth reported different levels of support from various close individuals (e.g., neighbor, teacher, parent/adult support) depending on their status of placement (i.e., in-home vs. out-of-home placement) and if they had been identified as academically at-risk, which is often the case among youth in foster care. However, in general, the association between social support, and the different sources of support, on academic functioning of youth in foster care is unclear, as very few studies have examined specific sources of social support among youth in foster care. It cannot be assumed that the relation between social support and academic functioning (both performance and behavioral health in school) will be the same for youth in foster care as compared to youth in the general population. Moreover, it is especially necessary to determine the importance of specific sources of social support given youth in foster care, by definition, experience changes in their social network and often lack continuity of support (Jones, 2014). Thus, the current study seeks to address this gap in understanding by examining several different sources of social support simultaneously in relation to both academic grades and behavioral wellbeing in school.

Further, this study sought to determine the role of social support while also accounting for the impact of another unique environmental factor that influences the academic wellbeing of youth in foster care – transitions while in care. More specifically, this includes both foster placement and school placement changes. Youth’s environments often lack stability while in care, as youth can experience multiple changes in their living situation or school while involved in the child welfare system (e.g., McGuire et al., 2018; Rubin et al., 2004). Although efforts have been made to prioritize youth sustained placement at their same home or school when in foster care (e.g., Fostering Connections to Success and Increasing Adoptions Act, 2008), this is not always possible and may result in negative academic outcomes. Several studies have documented an association between multiple home or school changes and below average performance in academic areas such as reading, writing, and mathematics amongst youth in foster care (e.g., Sullivan et al., 2010; Zima et al., 2000). Not only can these types of transitions have an influence directly on functioning (e.g., McGuire et al., 2018), but transitions in care and school can disrupt the typical academic development for youth in foster care due to missed educational content or a need to repeat grades (e.g., Zetlin & Weinberg, 2004). Therefore, the current study sought to determine how social support influences academic functioning while also accounting for the unique aspects of other notable environmental factors among youth in foster care that can influence academic performance and behavior.

Current Study

To build on the existing literature on academic functioning among youth in foster care, the current study examined the role of varying sources of perceived social support, specifically, the degree to which children perceived approval from a variety of potential support sources, in relation to academic performance and behavioral health in school. This included use of a longitudinal research design with structural equation modeling to simultaneously examine the explanatory power of several sources of social support (i.e., parents, teachers, classmates, and friends), across a variety of indicators of academic functioning, including grades from multiple subjects (i.e., math, language arts) and measures of both negative (i.e., internalizing, externalizing, school problems) and positive (i.e., adaptive skills) indicators of behavioral health. Additionally, the current study took into account unique aspects of the foster care experience by controlling for placement and school changes.

Given findings from the general literature about the role of teacher social support in relation to academic functioning, it was hypothesized that teacher social support would be positively associated with both the academic performance outcomes (i.e., math and language arts grades) and each of the four behavioral outcomes (i.e., externalizing problems, internalizing problems, school problems, and adaptive skills), and that teacher social support would have the strongest association with these outcomes, as compared to the other forms of social support. It was also hypothesized that none of the other social support source variables would be associated with the academic functioning outcomes. Further, given previous evidence demonstrating a potential negative influence from placement and school changes on academic functioning, it was also hypothesized that both placement and school changes would be negatively, significantly associated with both academic performance outcomes. No specific hypothesizes were made regarding the influence of placement and school changes on academic behavioral health outcomes.

Methods

Participants

Participants were 257 youth (M = 13.55 years, SD = 2.83 years), as well as their primary caregivers and teachers, recruited as part of a larger study of youth in foster care (Studying Pathways to Adjustment and Resilience in Kids; SPARK). The primary objective of the SPARK project was to examine potential factors associated with risk and resilience in relation to social, developmental, and emotional functioning of youth in foster care over time. Only youth who participated at all three time points (each three months apart) and who had a teacher complete measures of functioning at each time point were included in the current study since study aims required data from all three time points (Total SPARK sample enrolled at the Time One = 495). In cases where youth missed a time point, primary reasons for youth missing time points included: families not returning for data collection in the project’s 3-month window timeframe, children moving to different placements or leaving care, and loss of contact with families. Exclusion criteria for youth upon enrollment in the current study included caregiver report of (1) an Autism Spectrum Disorder diagnosis, (2) psychotic symptoms, (3) if youth were a non-native English speaker, (4) youth IQ below 70, or (5) if youth had not been residing in their current foster care placement for at least 30 days prior to baseline data collection (mean length of time in placement at baseline was ~90 days). This inclusion and exclusion criteria were used because of the demands of the larger project’s data collection procedures, as well as the study measures not being validated with youth who may not be native English speakers or who demonstrate cognitive challenges. Please see Jackson, Gabrielli, Tunno, and Hambrick (2012) for more information on recruitment procedures.

Youth were primarily male (54%; 46% female), and approximately 62% of youth resided in congregate care (e.g., residential facilities, group home settings) at the time of data collection, with the remaining 38% residing in traditional foster home settings (including both kinship and non-kinship foster home care). Most youth participants identified as African American (51%), followed by 31% White, 11% multiracial, and 7% other (see Table 1 for sample descriptive, as well as mean scores for the variables of interest described in the Measures section). Primary caregivers mostly consisted of foster care mothers (42.86%), followed by residential staff/case workers (29.76%), biological relatives (10.71%), foster care fathers (9.92%), and other (6.74%). Teachers consisted of 152 middle school and high school teachers from 24 different school district across the Midwestern county in which youth resided during data collection.

Table 1.

Sample descriptive and outcome scores for study sample.

Variable	Mean (SD)/Percentage
Gender (female)	45.6%
Ethnicity
Black	51%
White	30.9%
Multiracial	10.8%
Other race	7.2%
School Change	23.6%
Age	13.55 (2.83)
Placement Changes	9.59 (6.55)
Social Support
Parent Social Support	3.31 (0.74)
Teacher Social Support	3.41 (0.63)
Friend Social Support	3.42 (0.67)
Classmate Social Support	3.30 (0.65)
Academic Functioning
Language Arts Grade	3.27 (1.18)
Math Grade	3.71 (1.34)
BASC-TBS Externalizing	67.16 (15.97)
BASC-TRS Internalizing	68.82 (16.15)
BASC-TBS School Problems	61.25 (11.33)
BASC-TBS Adaptive	48.20 (8.62)

N = 257.

Measures

Social Support.

The Social Support Scale for Children (SSSC; Harter, 2012) is a 24-item measure that assesses for perceived parent¹, teacher, friend, and classmate social support. Youth report on aspects of perceived approval from these four support sources, such as perceived inclusiveness, understanding, fairness, time together, and helpfulness. Responses were based on two-part questions where youth needed to select (a) which statement applied most to them, and (b) how true the statement that they chose is of them (i.e., “really true for me” vs “sort of true for me”). For examples from each source (scoring shown in brackets), an item asking about parent social support includes: “Some kids have parents who care about their feelings [1] BUT Other kids have parents who don’t seem to care very much about their children’s feelings [0],” teacher social support includes: “Some kids don’t have a teacher who helps them to do their very best [0] BUT Other kids do have a teacher who helps them to do their very best [1],” friend social support includes: “Some kids have a close friend who really understand them [1] BUT Other kids don’t have a close friend who understands them [0],” and classmate social support includes: “Some kids have classmates who pay attention to what they say [1] BUT other kids have classmates who usually don’t pay attention to what they say [0].”² After selecting which part of the sentence represents them the most, youth would then rate whether the selected statement is “Sort of true of me” (scored as 0) or “Really true of me” (scored a 1). Responses were then totaled on a four-point Likert scale according to responses from both parts of the question, with higher scores for each of the four sources of social support on the SSSC indicating higher levels of perceived support from that source of support. Mean scores were then calculated for each of the four sources of social support by taking the mean score across all questions for each source of social support, which could range from 1.00 to 4.00. Means of all sources of support ranged from 3.30 to 3.42, consistent with means from other studies that indicate that means are typically negatively skewed and between approximately 3.00 and 3.50 (Lipski et al., 2013). The four mean scores for the parent, teacher, friend, and classmate subscales from Time 1 were included in the analyses as predictors of academic functioning at Time 3 (please see Table 1 for sample mean scores).

In previous studies with both school age and adolescent youth, the SSSC has demonstrated minimally acceptable to acceptable internal consistency (α=.72 to .81), acceptable test-retest reliability (r’s=.45 to .76), and acceptable internal validity for use in measuring social support across multiple sources of support (e.g., Harter, 2012; Lipski et al., 2014). The measure has also demonstrated acceptable convergent validity through correlations with the BASC-2 Relationship with Parents, Attitude to Teachers, and Interpersonal Relationships subscales (r’s=.42-.71; Lipski et al., 2014). In addition, there is support for the four-factor structure of the measure (Lipski et al., 2013). Internal reliability for each of the four subscales of the SSSC (as measured by Cronbach’s alpha coefficient [denoted as α]) in the current study ranged from acceptable to questionable: parent subscale α = .76, teacher subscale α = .69, friend subscale α = .72, and classmate subscale α = .66. As indication of internal validity, the intercorrelations between the subscale scores were in the moderate range (Table 2). Reliability estimates suggest there may be room for improvement of the SSSC scale within the foster sample, but for the purposes of this study, item-intercorrelations suggest scale scores provide adequate summaries of shared variance across items. While these alpha coefficients are low, there is some precedence for minimally adequate internal consistency cited in the literature on the SSSC (e.g., Kinard, 1995; Lipski et al., 2013).

Table 2.

Correlations between study variables.

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. Math Grade	1
2. Language Arts Grade	0.51*	1
3. Adaptive Skills	0.10	0.38*	1
4. School Problems	−0.25*	−0.38*	−0.85*	1
5. Externalizing Concerns	−0.16*	−0.32*	−0.78*	0.72*	1
6. Internalizing Concerns	0.01	−0.13*	−0.56*	0.56*	0.55*	1
7. Age	−0.09	0.00	0.00	0.14	0.13	0.10	1
8. Gender	−0.03	−0.22*	−0.39*	0.34*	0.30*	0.09	0.05	1
9. School Changes	0.02	−0.05	0.05	0.00	0.04	−0.10	−0.10	0.06	1
10. Placement Changes	−0.01	−0.07	−0.16	0.21	0.24	0.19	0.16	−0.03	0.04	1
11. Classmate Social Support	0.08	0.09	0.04	−0.11	0.06	−0.22	0.02	0.06	0.07	0.02	1
12. Teacher Social Support	0.25	0.13	0.18	−0.14	−0.14	−0.02	−0.05	−0.02	−0.10	−0.02	0.28	1
13. Friend Social Support	0.11	0.05	0.09	−0.07	−0.09	−0.03	0.13	−0.03	0.01	−0.12	0.54*	0.33*	1
14. Parent Social Support	0.12	0.06	−0.16	0.02	−0.04	0.16	−0.15	0.15	−0.05	−0.13	0.24	0.30	0.25	1

N = 257.

^*p < .05.

Behavioral Health in School.

The Behavior Assessment System for Children, Second Edition, Teacher Report (BASC-2-TRS; Reynolds & Kamphaus, 2004) was administered to assess behavioral (including social, emotional, and behavioral) health in school through teacher report. The BASC-2-TRS is a 160-item measure for ages 8 to 12 and a 150-item measure for ages 12 to 18. Responses were provided on a 4-point Likert scale ranging from 0 (Never) to 3 (Almost Always). For the purposes of this study, the T-scores for the Externalizing Problems, Internalizing Problems, School Problems, and Adaptive Skills composite scales were included from teachers’ reports at the final project time point (Time 3). The Internalizing Problems composite scale was comprised of the Anxiety, Depression, and Somatization subscales. The Externalizing Problems composite scale was comprised of the Hyperactivity, Aggression, and Conduct Problems subscales. The School Problems composite scale was calculated from items related to Attention and Learning Problems, and the Adaptive Skills composite scale consists of Adaptability, Social Skills, Leadership, Study Skills, and Functional Communication subscales. For each scale, total aggregate scores were converted to T-scores, and the T-scores were for each composite scale were included in data analysis.

In previous studies, the BASC-2-TRS has demonstrated satisfactory reliability (e.g., test– retest reliability, interrater reliability), as well as validity across several different indicators (e.g., construct validity, concurrent validity; Kamphaus, 2015; Reynolds & Kamphaus, 2004). The BASC-2-TRS demonstrated satisfactory internal reliability (as measures by Cronbach’s alpha coefficient [α]) across all scales for both the child and adolescent forms: Internalizing Problems-child form α = .91, adolescent form α = .89, Externalizing Problems- child form α = .98, adolescent form α = .96, School Problems- child form α = .75, adolescent form α = .87, and Adaptive Skills- child form α = .92, adolescent form α = .86.

Performance in School.

School grades were obtained from youth academic records (i.e., report cards). School administrators provided report cards during the semesters in which the participant completed data collection. School grades have been shown to be a valid predictor of other measures of academic achievement and cognitive ability, as well as a number of education outcomes at both pre- and post-secondary education levels (Brookhart et al., 2016; Geiser & Santelices, 2007). For example, across nationally representative studies, school grades tend to show a moderate to strong correlation with other more standardized measures of academic achievement and progress (Brookhart et al., 2016).

Grades reported at Time 3 were included in the analyses. The current study used two primary academic subjects as measures of performance in school: math (e.g., algebra, geometry) and language arts (e.g., communication arts, English, reading, and writing). These subjects represent primary academic curriculum youth tend to be enrolled in, and these subjects are the most often examined areas of academic performance across previous literature on the academic functioning of youth in foster care (e.g., Luke & O’Higgins, 2018). Additionally, these subjects allowed for a more similar grouping and summary across classes, as opposed to other forms of classes such as electives (e.g., art, Spanish). Letter grades from each area were converted into a continuous scale ranging from 1 (letter grade of F, or failing expectations) to 5 (letter grade of A, or exceeding expectations). If youth were enrolled in multiple classes for a given subject during data collection (e.g., enrolled in both a reading and a communication arts class), an average grade was calculated for that subject. Given the small number of youths in special education classes in the sample (likely due to exclusionary criteria), youth with grades from special education classes (including honors classes) were excluded from analysis using listwise deletion (n = 7).

Demographics.

Foster child demographic information was provided by the primary foster caregiver, including child age and gender. Given observed differences in the demonstration of behavioral concerns and academic performance across ages in childhood and genders (e.g., Hyde et al., 1990; Keiley et al., 2000), these variables were included in the model as covariates. This information was collected at Time 1.

School Changes.

Caregivers reported on the child’s current school placement at each data collection appointment. School changes was coded “yes” if school change occurred between Time 1 and Time 3 due to a non-normative transition (e.g., middle to high school was considered a normative transition and was not coded as a school change), or “no” if children were in the same school throughout the study.

Placement Changes.

Number of placement changes for each child was obtained from a placement database provided to the research team by the participating child welfare administration. Number of placement changes over the child’s lifetime was calculated by aggregating the total number of placements for each child.

Procedure

The Institutional Review Board from the authors’ university and the state’s Department of Social Services Review Board approved all study methods and measures. Participants were informed about the study using phone calls, flyers, advertisements in newsletters and listservs, and through referrals from other SPARK participants. Potential participants contacted the SPARK research team where they received more information about the study and were asked to complete a screener of eligibility. See Jackson et al. (2012) for a more detailed description of the recruitment process. Those who met inclusion criteria then scheduled a data collection appointment time. Data collection appointments took approximately three hours, and the same process was followed at each time point. Participants and the research team met in quiet and comfortable community locations (e.g., private room at library) easily accessible for the families. Although prior consent for youth to participate was obtained from the Division of Family Services and the Chief Judge of the Circuit Court, foster caregivers provided informed consent and youth provided informed assent at each time point.

After obtaining informed consent and informed assent, youth participants completed the study questionnaire. Ordering of survey items was counterbalanced. The questionnaire included completing the demographics and social support questionnaires using an audio-computer assisted self-interview (ACASI) program on a laptop computer. The ACASI is a user-friendly computer program that administers questions to participants both visually on screen and by verbally reading the questions aloud through headphones. ACASI use helps ensure participants with low reading, visual, or auditory abilities can answer questions through a private and confidential medium. Primary caregivers also completed the demographics measure using the ACASI system. Participants were provided with regular breaks and snacks while completing the study questionnaires. Following data collection, graduate student research assistants began an extensive debriefing process. The first part of the debriefing included assessing youth participants’ potential suicidal ideation and current abuse. Research assistants were specifically notified by the ACASI system to ask about possible current suicide and abuse if youth or the youth’s primary caregiver endorsed these question items on the ACASI. The second part of debriefing included checking for any significant changes in participants’ mood resulting from taking part in the study activities. Although findings suggest that asking youth about maltreatment history does not upset them significantly (Finkelhor et al., 2014), special precautions were taken to ensure participants were not distressed from research participation. This included precautions such as allowing children to take as many breaks as needed with continuous monitoring for emotional changes during completion of the survey, having research staff check on participants every 15–20 minutes, and comparing a pre- and post-mood survey rating for any changes in mood. Following the debriefing process, both the adult and youth participants received monetary compensation, which was $60 for the parent and $20 for the youth at the first time point, and the amount increased by $10 at each follow up time point. Follow up phone calls were made to each participant within 24 hours by graduate student research assistants to check for potential distress and to answer any questions they may have had about the study.

All area school districts were contacted in advance to obtain approval for collection of teacher reports on their students participating in the SPARK Project. Permission to access school records of the youth in the study was provided by the state social service agency, who was their legal guardian. In addition to filling out questionnaires, participants and caregivers provided information about what school the youth attended and what teacher they believed knew the child best so that their teacher could be contacted to provide information about youth’s academic functioning and behavior. The SPARK research team obtained the teacher’s email by contacting the school or by using online school directories. After confirming the identify of a teacher of the participant, the teacher was sent a link via email and asked to complete an electronic survey containing the BASC-2-TRS. All youth were known to the responding teacher for at least 30 days (in accordance with the instructions in the BASC-2-TRS) in order for teachers to complete the survey about a participant. Each teacher received $20 compensation for participation. Staff representing the youth’s school were also asked to send a current copy of the participant’s grade report card for all subjects.

Analytic Strategy

To evaluate the study’s hypotheses regarding the associations of sources of social support, as well as placement and school changes, with academic functioning (both school performance and behavioral health), structural equation modeling (SEM) with maximum likelihood estimation was employed to test these associations. This allowed for concurrent regression tests with multiple outcomes and predictors simultaneously. In a single model, all four sources of social support were included as predictors, which included the mean social support scores from teacher, parent, friend, and classmate. Age, gender, number of placement changes, and recent school change (y/n) were included in the models as covariates. All outcomes of interest were examined in the single SEM model, which included the two performance measures (i.e., math and language arts grades) and four behavioral health measures (Externalizing Problems, Internalizing Problems, School Problems, and Adaptive Skills). All outcome variables for assessing academic performance and behavioral functioning were from the final time point (Time 3) in the study, whereas all the predictor measures were reported on at Time 1 in the study, except the occurrence of a recent school change, which was measured between study time points prior to the last time point. Full information maximum likelihood was employed to manage missing data for participants with missingness (ranging from 1 to 8% across all variables) under the assumption of data missing completely at random (Kline, 2015; Raykov, 2005). All models were fit using R software (R Core Team, 2014). All covariances between outcome scores for the performance and behavioral health outcomes were modeled, as well as all the covariances between each of the sources of social support subscales (Figure 1).

Fig. 1.

Model for School Performance and Behaviors. The non-significant pathways between the factors and the outcomes variables are not shown to improve clarity. Significant pathway estimates (p < .05) shown with bold line. Unstandardized regression estimates are shown above each significant pathway.

Results

Means and standard deviations or percentages for the variables of interest are reported in Table 1. Across the three-time points (approximately six to nine months), 24% of youth had non-normative school transitions, and the average number of lifetime placement changes at Time 1 was approximately nine. Regarding academic grades, youth appeared to be performing at approximately a C average level at Time 3. Further, the mean scores for the behavioral health outcomes were all below but close to the clinically significant cutoff (i.e., T-score < 70) for the BASC-2-TRS. There were similar levels of social support overall reported across youth in the current study. The highest rated perceived social support source on average was from friends, followed by teachers, parents, and then classmates. The correlations between each of the study variables of interest are presented in Table 2.

Relation Between Social Support and Academic Functioning

The estimated model is shown in Figure 1. In the model, all modeled covariances were significantly correlated with each other at the p < .01 level, with the exception of the covariance between math grades-adaptive functioning (p = .14), math grades-internalizing symptoms (p = .82), and math grades-externalizing symptoms (p = .15). Further, there was a non-significant covariance between internalizing symptoms and language arts grades (p = .17). The model demonstrated adequate fit to the data (χ2 _{(22, n = 257)} = 44.84, p < .01, RMSEA _{(.03- .09)} = .06, CFI = .97, SRMR = .04).

For math grades, teacher social support was the only significant predictor (Β = .48; p < .01), suggesting that youth with greater perceived social support from teachers tended to obtain higher math grades than those youth with lower perceived support from their teacher. For language arts grades, gender was the only significant predictor (Β = −.54; p < .01), which suggested that on average, males tended to receive lower grades in language arts as compared to females. The following variables were not associated with grade outcomes: age, parent social support, friend social support, classmate social support, placement changes, and school changes.

In relation to adaptive skills, there was a significant positive association for perceived teacher social support (Β = 2.58; p < .01). This suggests that youth with more perceived teacher social support tend to demonstrate more adaptive skills. There was a significant negative association between adaptive skills and both gender (Β = −6.171; p < .01) and placement changes (Β = −.26; p = .01). This suggests that males (compared to females) and those youth with a greater number of placement changes tend to demonstrate fewer teacher reported adaptive skills. A similar but opposite pattern was observed for the school problems outcome, such that there was a significant, negative association for perceived teacher social support (Β = −2.80; p = .03), and a significant, positive association for gender (Β = 7.67; p < .01) and placement changes (Β = .33; p = .01). Youth with more perceived teacher social support tended to demonstrate less school problems (i.e., attention and learning problems), as did females and those with fewer placement changes.

For externalizing concerns, there were two significant predictors, gender (Β = 9.71; p < .01) and placement changes (Β = .50; p = .02). These findings suggest that males and those youth with a high number of placement changes tended to demonstrate more externalizing difficulties on average as compared to females and those with fewer placement changes, respectively. For internalizing concerns, perceived classmate social support was the only significantly negative predictive predictor (Β = −5.83; p = .02), indicating that on average, youth with more perceived social support from their classmates tended to demonstrate fewer internalizing difficulties in the school setting than those who did not perceive social support from their classmates. There were also significant positive associations between internalizing problems and placement changes (Β = .53; p = .03) and perceived parent social support (Β = 4.12; p = .02). These findings suggest that youth with more placement changes and more perceived parent social support tended to demonstrate more internalizing symptoms in the school setting as reported by teachers. The following variables were not associated with any of the teacher reported behavioral outcomes: age, friend social support. and school changes.

Discussion

Youth in foster care demonstrate more difficulty than their non-foster care peers in academic settings (Stone, 2007; Trout et al., 2008). One known modifiable environmental factor in the general population that promotes academic functioning, including both performance and behavioral health in school, is perceived social support. However, there is a dearth of research examining how social support may influence academic functioning for youth in foster care. The current study sought to expand the knowledge base regarding which sources of social support may contribute to positive functioning in school, while taking into account factors associated with the foster care experience (e.g., placement changes, school changes) that may influence both academic functioning and social support.

Overall, youth reported experiencing social support frequently across all four sources of social support studied, which appears to be in line with reporting behaviors in other samples of youth (i.e., similar mean scores across sources of support; e.g., Lipski et al., 2014). However, despite the presence of perceived social support across different sources, differences in how social support from each of these sources influence academic functioning emerged. In line with study hypotheses, results revealed that perceived teacher social support was associated with performance in math (i.e., math grades), and significant associations were also observed between teacher social support and behavioral health in school. Previous literature has demonstrated the importance of teacher social support for successful academic performance and positive behavioral in the general population (e.g., Pan et al., 2017). The results of the current study expanded on these findings by showing that, even among youth in foster care who often experience multiple changes in their social and school structure, perceived social support from teachers still may significantly contribute to academic functioning. Further, even when taking into account other sources of social support, perceived teacher social support was still a predictor of behavioral health and performance in school over and above the other sources of social support. For youth in foster care, teacher social support may be more influential than social support from other close individuals (e.g., parents/caregivers, friends) in a school setting because there may be more continuity (i.e., more predictable and structured interactions) with teachers (Jones, 2014). Thus, the formality that exists within a student-teacher relationship compared to relationships with parents or peers, may create a safe, less threatening, accepting, and enriching learning environment for youth in foster care (Chen et al., 2003; Farruggia et al., 2006).

Compared to the other sources of social support, teacher social support overall appears to be most important in relation to academic functioning given the number and direction of observed findings between perceived teacher social support and both behavioral health and performance in school. For example, with the exception of gender, perceived teacher social support was the only predictor found to be significantly associated with both academic performance and behavioral health in school. Although teacher social support provided the most consistent and stable significant associations with academic functioning in this sample, this is not to say that social support from other individuals is not important for youth in foster care. However, for youth in foster care, social support may be most predictive for specific aspects of functioning depending on the match between the source of the support (e.g., teacher, parent) and related environment (e.g., school, home; Jones, 2014). Thus, when closely comparing social support of foster care youth across different sources and environments, teacher social support may be the most influential in a school setting, whereas parent social support may for example be most influential in the home setting.

There were other notable findings with regard to social support and academic functioning. In the current study, perceived social support from classmates was negatively associated with internalizing symptoms in school, such that youth who reported more social support from those around them at school tended to have teachers report fewer internalizing concerns. These findings extend on several other studies in the general population, with at-risk youth demonstrating the positive impact social support from peers may have on behavioral health (e.g., Ezzel et al., 2000; Rueger et al., 2010). Perceived social support from classmates who are close in age to the youth may promote aspects of wellbeing, such as positive self-esteem and skill mastery, which can either directly influence internalizing symptoms or provide a buffer against life stressors that may increase internalizing symptoms (e.g., Lipschitz-Elhawi & Itzhaky, 2005). Moreover, as youth age there likely is a shift in strength of support and reliance on parents to classmates and peers for one’s emotional wellbeing (Nickerson & Nagle, 2005; Rueger et al., 2010; Van Beest & Baerveldt, 1999). Given that the mean age of participants in this study was about 13 years old, this could be true for individuals in this sample. These findings may suggest that among youth at-risk for low social support from parents, such as youth in foster care, classmate or peer social support may be especially important for mental health functioning compared to other sources of support.

It is worth noting the unexpected finding that perceived parent social support was positively associated with internalizing symptoms observed in school. In general, the support of parents among youth in foster care and the general population is often found to play a protective role for youth functioning, in that youth who perceive greater support from their parents tend to report fewer internalizing concerns compared to those who report lower parent support (Hombrados-Mendieta et al., 2012). While it is the case that some studies on youth in foster care and the general population have reported no association between parent or caregiver social support and internalizing symptoms (e.g., Rueger et al., 2010; Salazar et al., 2011), it has not been reported that there is a positive association between parent social support and internalizing symptoms. One explanation is that youth who experience a high degree of internalizing symptoms might solicit a higher degree of parent or caregiver social support in comparison to youth who do not experience problematic internalizing symptoms (Hughes & Gullone, 2008). Further evidence for this possibility comes from items on the SSSC (Harter, 1985), which ask about how parents respond to their child’s feelings and problems. Another possible explanation for the link between internalizing problems and high levels of perceived parental social support may be that youth have concerns about believing they are reliant on their parents. For example, one qualitative study on adolescent youth in foster care reported that adolescents experienced more anxiety when believing that they were dependent on family and others (Cunningham & Diversi, 2013). However, given the little research available on specific sources of social support in relation to behavioral health in school among youth in foster care, more research is needed to identify why the observed finding may have occurred.

In addition to findings on social support, results from the current study build on previous findings regarding gender and academic functioning. Gender, in addition to perceived teacher social support, was the only predictor significantly associated with both academic performance and behavior in school. This is consistent with previous research suggesting that males in foster care tend to demonstrate higher rates of behavioral problems (e.g., Kaiser et al., 2000; Rose & Rudolph, 2006) and receive lower grades in language arts classes (Voyer & Voyer, 2014), as compared to females in foster care. Moreover, study findings also highlight the importance of a youth’s living situation on their ability to function well in school. Even more than teacher social support, placement changes was the most consistently influential variable associated with academic functioning in foster care youth. In the current sample, the data suggested a higher frequency of placement changes was associated with more negative behavioral health in school (i.e., fewer adaptive skills, and more internalizing, externalizing and school problem symptoms) among youth in foster care. These findings are consistent with previous studies that have demonstrated an association between changes in placement and poor behavioral health in school for youth in foster care (e.g., Zima et al., 2000). The robustness of the influence of placement changes on academic functioning in youth in foster care is also not surprising because there are multiple reasons how placement changes may influence academic functioning. For example, due to the shift in home and school placements, foster youth may fall behind academically because they miss fundamental academic concepts taught in class during their transitions (Zorc et al., 2013). Some youth also experience delays in receiving credits from previous schools or poor documentation of previous school records, increasing the risk of class and grade repetition (Zetlin et al., 2012).

Interestingly, and in contrast to previous literature (e.g., Martínez et al., 2011; Sullivan et al., 2010), changes in school placement were not associated with performance or behavioral health in this sample of foster care youth. One possible explanation is that changing schools may have provided youth in foster care with different opportunities to improve their self-competence and establish social networks in a new school environment (Pears et al., 2012; Weiss & Bearman, 2007). In turn, this opportunity to make new friends, meet new classmates, and/or feel more confident in their abilities could be related to fewer problems with behavioral health in general (Newman et al., 2007; Rosario et al., 2008). The finding regarding a positive association between classmate social support and internalizing concerns may further endorse this possibility. In addition, the duration of the present study (approximately six months) may not have been long enough to fully identify participant experiences of school disruptions. A better measure of the number of non-normative school disruptions across a child’s entire academic history may be a more informative covariate in future studies.

Taken together, these results speak to the multifaceted and complex relation when considering multiple sources of social support, the experiences of youth in foster care, and academic functioning. Overall, results suggest environmental factors, such as social support from a variety of different individuals and placement changes, appear to be strongly associated with school functioning. This pattern held constant for both academic performance and behavior. The current findings are consistent with previous literature that suggests that social and environmental factors may still have a strong, direct influence school functioning (Dalton et al., 2007; Elias & Haynes, 2008), despite the fact that youth in foster care tend to live in dynamic and changing social environments.

Study findings should be considered in light of some limitations. Although school grades can provide a more objective measure of academic performance compared to self- or teacher-report of student academic progress (e.g., Seyfried et al., 2000) and have been shown to be closely associated with other more standardized measures of academic achievement (Brookhart et al., 2016), curriculums and scoring procedures often vary greatly by school (Porter et al., 2011). While not uncommon to measure academic performance using school grades (e.g., Brookhart et al., 2016; O’Higgins et al., 2017), it is possible that nationally normed, standardized tests of academic achievement could reduce school-specific grading biases, such as grading biases among minority students (e.g., Walton & Spencer, 2009). However, it is also important to consider the potential biases in standardized testing (e.g., Warne et al., 2014). Second, the current study was comprised of youth in foster care from the same Midwestern county, and there was a larger portion of youth from congregate care settings (e.g., residential facilities, group home settings) in the current study from this county compared to the state and national averages (e.g., Missouri Department of Social Services, 2014). Together, this may limit the generalizability of the study’s findings. For example, it may be the case certain specifics of the foster care system within this county (e.g., strategies for placement changes, percentage of youth in congregate care), as well as possible differences in certain variables within the study (e.g., mental health symptoms) that have been previously shown to vary between youth in foster care families and congregate care settings (e.g., Okpych & Courtney, 2018), may have uniquely contributed to study findings. Further, the current study did not recruit youth from foster care settings specific for youth with known intellectual or developmental disabilities. Thus, this may have excluded youth with a high probability of placement in special education and higher rates of lower academic achievement (e.g., Scherr, 2007).

Future Directions and Conclusions

Several recommendations for future research and intervention are noted. First, it is essential to emphasize the importance of teacher social support on the performance and behavioral health in school for youth in foster care. One area of future research should be to more closely examine various mechanistic aspects of teacher social support. Research is needed to identify specific mechanisms of social support (e.g., emotional support, informational support) and the behaviors demonstrating these types of support that teachers engage in that may help explain what aspects of teacher social support influence youth in care to be successful in school. However, this may not only apply to teacher social support and future studies should continue to explore social support as a multidimensional construct that includes different sources of social support and the different types of support that can be provided. For example, while the current study focused on four sources of social support, there are a several other sources of social support that may influence academic performance and behavioral health in school (e.g., school counselor, sport’s coach, social worker or therapist). Researchers should not only be focused on social supports identified a priori in validated survey measures, but also consider asking youth in foster care to identity their most relevant sources of social support in an open-ended fashion.

Furthermore, while the current study focused on the direct model of social support given the limited literature to date on youth in foster care and social support as it relates to academic functioning, it is necessary in future research to explore the indirect, buffering effects of social support on youth in foster care. Evidence from studies with youth in foster care and in the general population suggests that social support may attenuate the influence of other factors in a child’s social environment that could negatively influence functioning (e.g., Parkes & Sweeting, 2018; Salazar et al., 2011). As empirical evidence on the influence of specific sources of social support among youth in foster care continues to grow, it will be necessary to compare and contrast models of direct and indirect effects from social support to better understand how social support may be most influential among these youth, as well as possible transactional models of social support with student-teacher relationships (e.g., Sutherland & Oswald, 2005).

Following the establishment of clear evidence base on the influence of social support on academic functioning for youth in foster care, empirical findings could then support the developmental of clinical or intervention programs in this area. This appears to continue to be a gap in the literature for youth in foster care, as research and practice have largely overlooked direct intervention for academic functioning of youth in foster care. For example, reviews focused on interventions that seek to address concerns with academic functioning among youth in foster care tend to only find slightly over 10 studies that meet inclusion criteria (Evans et al., 2017; Forsman & Vinnerljung, 2012). Future research should therefore address how providing additional training and resources for teachers may result in even better academic functioning for youth in foster care. This may include options such as providing teachers with specific training on how to establish a relationship, sustain engagement, and motivate youth in foster care who demonstrate problematic school behaviors, in addition to addressing logistical or structural barriers (e.g., time to work one-on-one with a student) to allow a teacher to properly implement these skills or strategies (Forsman & Vinnerljung, 2012; Pecora, 2012; Weinberg et al., 2009). This could be especially beneficial for teachers who may feel unprepared to work with youth in foster care, perhaps given their frequently complex history of trauma exposure (Pickens & Tschopp, 2017; Zetlin et al., 2012). For example, Zeltin et al. (2012) found new school teachers working with youth in foster care felt they were unprepared to handle their behavioral and academic problems, had no additional assistance specifically for this group of students, and cited needing more training and education on this population. Additionally, information on social support could also be used to help inform individual or group tutoring programs for youth in foster care (e.g., Flynn et al., 2012; Harper & Schmidt, 2016). This might include ensuring important sources of support are involved in the tutoring process (e.g., parent, classroom teachers if the youth’s teacher is not the tutor) or if different types of social support that may increase the effectiveness of these interventional programs. Moreover, methods for increasing collaboration between schools and child welfare should be further researched. This might include policy level changes that allow for information under specific guidelines to be shared across agencies, efficacy of interdisciplinary trainings, and working to give agencies greater access to school services such as tutoring and academic mentoring programs (Altshuler, 2003).

Lastly, with regard to future directions in research on the academic functioning of youth in foster care, results from this study suggest that research inquiries should consider broader environmental level factors that may contribute to youth functioning when studying academic performance and behavior. For example, placement related factors were consistently associated with academic functioning, in addition to social support. Thus, these foster care specific factors (e.g., placement changes) should be examined to provide more clarity regarding which risk factors more significantly contribute to academic success for youth in foster care. Future research may also consider examining the circumstances associated with a placement or school change, youth’s perception of these changes, and the association of these factors with school functioning. For example, this might include examining how school changes may influence academic functioning if the change was the result of a placement change or it was determined that changing schools would be in the best interest of the child based on other individual or environmental factors (e.g., access to more educational services).

This may also include placing importance on educating school personal (including teachers and school administrators) on the legal and educational rights of youth in foster care (Weinberg et al., 2009). For example, this might include education on the requirement for child welfare and educational agencies to prioritize keeping youth at their school of origin and providing transportation for youth to this school if it is deemed to be in the best interest of the youth to not undergo a school change (e.g., Fostering Connections to Success and Increasing Adoptions Act, 2008). In conclusion, it is notable that teacher social support may play an important role for the academic functioning of youth in foster care. Many teachers may feel somewhat powerless to help children in foster care given all of the factors which encompass the life of a youth in foster care that remain outside of the teacher’s control. However, teachers can make a difference through the level of providing consistent social support as this appears to be a powerful agent of academic success for youth in the foster care system.

Highlights.

• Four unique sources of social support (teachers, parents, friends, and classmates) for youth in foster care were examined in relation to academic performance and behavioral health in school.

• Teacher social support was the only source of social support to be associated with both academic performance (e.g., math grades) and behavioral health (e.g., adaptive skills) outcomes.

• Results also suggest that placement changes can have a negative influence on behavioral health functioning in school for youth in foster care.

• Findings highlight the importance of examining environmental factors when considering what might contribute to the academic wellbeing of youth in foster care.