Are Sexual Minority Youth Overrepresented in Foster Care, Child Welfare, and Out-of-Home Placement? Findings from Nationally Representative Data

Abstract

Background

Preliminary evidence suggests that sexual minority (e.g. lesbian, gay, bisexual, and same-sex attracted) youth are overrepresented in child welfare services. Yet, no study to date has been able to test this hypothesis with national data.

Objective

Using a two-study design, we test whether sexual minority youth are overrepresented in child welfare, foster care, and out-of-home placement using nationally representative data from the United States.

Participants and setting

Study 1 data are from the National Longitudinal Study of Adolescent to Adult Health (n = 14,154; Mean age = 15.4). Study 2 data are from wave three of the National Survey of Child and Adolescent Well-Being II (n = 1,309; Mean age = 15.0).

Methods

For Study 1, we use adjusted logistic regression models to test differences in lifetime foster care involvement between sexual minority and heterosexual youth. In Study 2, we calculate a Disproportionality Representation Index (DRI) – a ratio of sample prevalence relative to the general population – to estimate whether sexual minority youth were overrepresented in child welfare and out-of-home care.

Results

Study 1 results indicate that sexual minority youth are nearly 2.5 times as likely as heterosexual youth to experience foster care placement (aOR = 2.43, 95% CI 1.40, 4.21, p = .002). Results from Study 2 show that sexual minority youth were largely overrepresented in child welfare services (DRI = 1.95 to 2.48) and out-of-home placement (DRI = 3.69 to 4.68).

Conclusions

Findings are the first to demonstrate sexual minority youth’s overrepresentation in child welfare, foster care, and out-of-home placement using nationally representative data and emphasizes the need for focused research on sexual minority youth involved in the child welfare system.

Child welfare provides temporary care through out-of-home placements, such as foster care, to youth who have been removed from their home due to abuse or neglect (Pecora, Whittaker, Maluccio, & Barth, 2000). The overrepresentation of marginalized youth in foster care is a pressing public health concern. Anecdotal and preliminary empirical evidence suggests that sexual minority youth (i.e., lesbian, gay, bisexual, or same-sex attracted youth) enter foster care at disproportionate rates, and receive worse treatment, compared to their heterosexual peers when involved in foster care systems (e.g., Dworsky & Hall, 2013; Estrada & Marksamer, 2006; Gallegos et al., 2011; Gilliam, 2003; Irvine & Canfield, 2015; Mallon, Aledort, Ferrera, 2002; Tarnai & Krebill-Prather, 2008; Wilson & Kastanis, 2015). A recent study, for example, found that 19% of a community sample of foster youth in Los Angeles County identified as LGBTQ (lesbian, gay, bisexual, transgender, or queer), nearly three times higher than the estimated 6–8% of LGBTQ youth in the general population (Wilson & Kastanis, 2015). Given findings in community-based samples that document the overrepresentation of sexual minority youth in child welfare systems (e.g., Dworsky & Hall, 2013; Wilson & Kastanis, 2015), the Administration for Children and Family Office of Planning, Research and Evaluation (Burwick, Gates, Baumgartner, & Friend, 2014) recently emphasized the need for representative and population-based estimates of sexual minority youth involved with child welfare to better understand their experiences and unique needs. To address this need, the current study used data from two nationally-representative data sources from the United States to examine whether sexual minority youth are more likely than heterosexual youth to be involved in child welfare, foster care, and out-of-home placement.

Sexual Orientation Disparities in Child Welfare

The link between sexual minority status and entry into child welfare services is not well documented in the empirical literature. There is, however, a growing body of literature which documents strained parental relationships for sexual minority youth (Bouris et al., 2010; Ryan, Huebner, Diaz, & Sanchez, 2009; Willoughby, Doty, & Malik, 2010). Compared to heterosexual youth, for example, sexual minority youth are more likely to experience psychological and physical abuse (Baams, 2018; Friedman et al., 2011) that may directly or indirectly contribute to experiences with child welfare and out-of-home placements (Mallon, Aledort, & Ferrera, 2002; Remlin, Cook, & Erney, 2017). More recent work also suggests that sexual minority youth are more likely to run away or be asked to leave their families of origin than heterosexual youth (Durso & Gates, 2012; Pearson, Thrane, & Wilkinson, 2017) and that these experiences are attributed to rejection related to their sexual orientation (Durso & Gates, 2012; Wilson & Kastanis, 2015). For example, sexual minority youth who disclose their sexual orientation are more likely to report victimization and rejection from family (D’Augelli, Herschberger, & Pilkington, 1998; D’Augelli, Grossman, & Starks, 2008; D’Augelli, Grossman, Starks, & Sinclair, 2010). These narratives are particularly evident when we consider the overwhelming numbers of sexual and gender minority youth who experience homelessness (Durso & Gates, 2012). Given that modern cohorts of youth are coming out at younger ages (Gates, 2017; Russell & Fish, 2016), contemporary sexual minority youth may experience rejection and pushout during early and middle adolescence, a time when they are still wholly dependent on care from their families.

Though state agencies do not systematically track the sexual orientation of youth in foster care, two empirical studies relying on community samples find that sexual minority youth are largely overrepresented in regional foster care systems. The first, conducted in Iowa, Wisconsin, and Illinois from 2002 to 2011, found that nearly 15% of youth transitioning out of foster care identified as lesbian, gay, or bisexual (Dworsky & Hall, 2013). The second study (Wilson & Kastanis, 2015) used a community-based sample of foster care-involved youth from Los Angeles County and found that more than 19% of these youth reported an LGBTQ identity – two to three times higher than what was expected given national estimates of sexual minority youth ages 12–18 in the Youth Risk Behavior Survey (YRBS; 7–8%, Kann et al., 2011), among adults 18 and older in the Gallup Poll (6.4%; Gates and Newport, 2013), or youth ages 12–18 in the YRBS for Los Angeles Unified School District (7.1%, LAUSD, 2013). Together, these two studies, though regionally dependent, clearly point toward a disproportionate presence of sexual minority youth in foster care.

Disparities in Experiences of Care

Preliminary evidence suggests that once in foster care and out-of-home placement, sexual minority youth are more likely to experience maltreatment than heterosexual youth (Mallon, 1998; Mallon et al., 2002; Wilson & Kastanis, 2015). For example, sexual minority youth in foster care report discrimination from foster parents, foster siblings, and foster care staff (Gilliam, 2003; Mallon, 1998, Mallon et al., 2002), and are more likely than their heterosexual peers to be hospitalized for emotional and physical reasons (Wilson & Kastanis, 2015). Sexual minority youth in foster care also report receiving fewer services and more placements than heterosexual youth (Gallegos et al., 2011; Mallon, 1998; Wilson & Kastanis, 2015). An unsafe and unaccepting environment can contribute to a lack of permanency in care for sexual minority foster youth. Sexual minority youth in foster are less likely to be reunified or adopted (CASA, 2009; Mallon et al., 2002), have a higher number of placements (Dworsky & Hall, 2013; Wilson & Kastanis, 2015), and are more likely to be placed in restrictive settings (Dworsky & Hall, 2013). If reunification with their family of origin is not a safe or viable option for sexual minority youth, staying in care may be a safer alternative. From a systems perspective, awareness of sexual minority-specific issues, such as entry into child welfare and child care systems, is needed to improve current training addressing the unique needs of sexual minority youth in child welfare, foster care, and out-of-home placement (Burwick et al., 2014).

The Current Study

Considering that sexual minority youth and youth involved in child welfare are independently at risk for poor mental and behavioral health outcomes relative to heterosexual (Marshal et al., 2008, 2011) and non-system involved youth (Braciszewksi & Stout, 2012) respectively, we might imagine that the child welfare and foster care experiences for sexual minority youth are unique and complex in ways that may compound risk for poor outcomes. This supposition is supported by previous work that demonstrates that sexual minority youth placed in state care experience maltreatment from foster family and service staff (Feinstein et al., 2001; Gilliam, 2003; Mallon, 1998; Wilson & Kastanis, 2015), as well as a higher number of housing transitions (Dworsky & Hall, 2013; Wilson & Kastanis, 2015). Though preliminary research notes sexual minority youth disparities in foster care placement within specific states or regions (Dworsky & Hall, 2013; Mallon, 1998; Wilson & Kastanis, 2015), the overrepresentation of sexual minority youth in child welfare has not been documented using nationally representative data (Burwick et al., 2014). In order to adequately study the overrepresentation of sexual minority youth in child welfare and out-of-home placement, population-based data that include measures of sexual identity, behavior, and attraction are necessary, as are surveys that adequately assess youth living situations and transitions (Burwick et al., 2014). Using two nationally representative studies, one general adolescent population sample and one from adolescents involved in child welfare, we assess whether sexual minority youth are more likely to be involved in child welfare, foster care, and out-of-home placement.

First, in Study 1, we use data from the National Longitudinal Study of Adolescent to Adult Health to estimate sexual orientation disparities in foster care placement. Specifically, we test whether youth reports of same-sex romantic attraction are associated with foster care involvement prior to the age of 18. We hypothesized that these sexual minority youth would be overrepresented in foster care (H1). Next, in Study 2, we use data from the National Survey of Child and Adolescent Well-Being II (NSCAW II) – a nationally representative longitudinal survey of children and families who have been investigated by Child and Protective Services – to estimate the degree to which youth reports of sexual identity and attraction are overrepresented in child welfare and out-of-home placement. We also use the NSCAW II to assess whether sexual minority youth were more likely to report out-of-home placement relative to heterosexual youth involved with Child and Protective Services, findings which may indicate differential movement into or out of the child welfare system. We hypothesized that sexual minority youth would be overrepresented in the child welfare sample as well as among those reporting out-of-home placement (H2a). We also anticipated that, among youth involved with Child Protective Services, sexual minority youth would be more likely than heterosexual youth to report advancement to out-of-home placement (H2b).

Study 1: National Population Estimates

For Study 1, we assessed the prevalence of foster care involvement among heterosexual and sexual minority youth using a nationally representative sample of adolescents.

Method

Data Source and Sample

Data are from The National Longitudinal Study of Adolescent to Adult Health (Add Health; Harris, 2009), which is one of the largest nationally representative, longitudinal studies following US adolescents into adulthood. Wave I was collected during the 1994–1995 school year from approximately 90,000 7^th-12 graders from 80 high schools who were selected to represent the US population at that time. From this school-based sample, 20,745 youth were selected to complete an in-home interview questionnaire. Wave 2 was collected a year after and included 14,738 of the original participants. Wave 3 was collected in 2001–2002 when participants were between the ages of 18–26 (n = 15,197) and Wave 4 was conducted in 2008 when participants were between the ages of 24–32 (n = 15,701). For the current study we include those youth who provide valid data for same-sex attraction at Wave 1 and foster care involvement at Wave 3. We also used available data in Wave 1 and Wave 2 to measure foster care involvement, although Wave 2 participation was not a required inclusion criterion for the current sample. Demographic characteristics of our final analytic sample (n = 14,154) are presented in Table 1.

Table 1.

Study 1: Sample Characteristics of the National Longitudinal Study of Adolescent to Adult Health (N = 14,154)

	n	%w	(95% CI)	% missing
Biological Sex				0.00
Male	6,678	50.78	(49.53, 52.02)
Female	7,476	49.22	(57.98, 50.47)
Race/Ethnicity				0.04
White	7,622	67.25	(61.34, 72.66)
Black	2,967	15.74	(12.07, 20.28)
Latino	2,131	10.95	(8.15, 14.55)
Other	1,428	6.06	(4.48, 8.14)
Public Assistance	1,025	8.85	(7.31, 10.66)	14.12
Parental Education				5.28
< High School	1,705	12.30	(10.31, 14.61)
High School	3,918	31.60	(29.28, 34.01)
Some college	2,816	21.22	(19.88, 22.62)
≥ College	4,967	34.88	(31.32, 38.62)
Lifetime Foster Care Involvement				0.00
No	13,789	97.86	(89.92, 91.60)
Yes	365	2.14	(1.81, 2.54)
Romantic Attraction				0.00
Exclusively other sex	13,324	81.48	(79.86, 82.99)
No attraction	678	12.26	(10.77, 13.91)
Any same-sex	152	6.27	(5.59, 7.02)

Note. M_age = 15.44, 95% Cl (15.20, 15.67) at Wave 1

Measures

Lifetime foster care involvement

Participants’ lifetime involvement with foster care was measured using three items. First, in Wave 1 and 2, youth provided a household roster of who they were living with at the time. If they indicated they were living with a parent, they could specify if that parent was a foster parent. Second, in Wave 3, young adults were asked, “Did you ever live in a foster home?” If participants indicated living with a foster parent during Wave 1 or Wave 2 or if they reported ever living in a foster home at Wave 3, they were coded as having lifetime foster care involvement = 1. Participants who did not live with a foster parent or in a foster home were coded as having no foster care involvement = 0.

Adolescent romantic attractions

In Wave 1, participants were asked two questions: “Have you ever had a romantic attraction to a female?” and “Have you ever had a romantic attraction to a male?” (no = 0, yes = 1). Items were recoded in conjunction with participants’ biological sex to reflect youth who indicated no attractions, exclusively different-sex attraction, and any same-sex attraction. Due to power constraints, this last group includes both youth with exclusive same-sex attraction and youth who report attraction to both boys and girls.

Analytic Strategy

We first used chi-square test of independence to test whether sexual minority youth, on the basis of same-sex attraction, were more likely than different-sex attracted youth to report lifetime foster care involvement. Next, we conducted adjusted logistic regression models to estimate the effect of youth sexual minority status on lifetime foster care involvement adjusting for other confounding factors of system involvement: biological sex (Add Health does not include measures of gender identity), race/ethnicity, receipt of public assistance, and parental education. All analyses were adjusted to account for the complex, longitudinal and clustered survey design of the Add Health data and weighted to provide population-based estimates that generalize findings to the overall US population using “svy” procedures in Stata. We conducted multiple imputation to account for missing data on covariates for adjusted logistic regression models.

Results

Similar to the overall sample in Add Health, 51% of our sample was male and majority White (67%), followed by 16% Black, 11% Latino, and 6% Other. Approximately 9% reported that their parents received public assistance as an adolescent. Nearly 35% of youth reported that their parents had a college degree or higher, followed by a high school degree (32%). Almost 6.3% of youth reported any same-sex attraction during adolescence and 2.14% of participants indicated foster care involvement.

Chi-square test of independence indicated a significant association between sexual minority status and lifetime foster care involvement, χ² (2, N = 14,154) = 6.36, p = .002 (see Table 2). Specifically, 4.31% of sexual minority youth reported foster care involvement compared to 1.86% heterosexual youth. Results from adjusted logistic regression models indicated that for sexual minority youth, the adjusted odds of reporting foster care involvement are 2.43 (95% CI [1.40, 4.21]), p = .002) times higher than the odds of heterosexual youth. Given a growing body of literature that documents elevated risk among bisexual adolescence, we also conducted a follow up logistic regression with exclusively same-sex attracted and youth attracted to both girls and boys separated. Results indicated that both exclusively same-sex attracted (aOR = 4.16, 95% CI [1.70, 10.14, p = .002) and bisexually-attracted youth (aOR = 2.02, 95% CI [1.07, 3.82], p = .028) had greater odds of foster care involvement than heterosexual youth, although the effect was stronger for exclusively same-sex attracted youth.

Table 2.

Study 1: Chi-square Analyses Testing Adolescent Foster Care Involvement and Same-Sex Attraction in the National Longitudinal Study of Adolescent to Adult Health (N = 14154)

	No Foster Care Involvement			Foster Care Involvement			χ2,	p-value
	n	%w	(95% CI)	n	%w	(95% CI)
No Attraction	1,560	97.10	(95.50, 98.14)	55	2.90	(1.86, 4.50)	6.36,	.002
Exclusively different-sex attraction	11,433	98.14	(97.74, 98.47)	308	1.86	(1.53, 2.26)
Any same sex attraction	796	95.69	(97.46, 98.19)	35	4.31	(2.71, 6.78)

Study 2: Overrepresentation in Child Welfare and Out-of-Home Placement

For Study 2, we use a US nationally representative sample of youth involved with Child Protective Services to assess the prevalence of heterosexual and sexual minority youth involved in child welfare involvement and out-of-home placement. We also test whether sexual minority youth were more likely than heterosexual youth to report out-of-home placement, among youth already involved in child welfare services.

Method

Data Source and Sample

NSCAW II is a Congressionally-mandated national, longitudinal study designed to collect data from children and families involved within the Child Protective Services. Wave 1 was collected in March 2008 and two subsequent waves of data were collected in 18 month intervals. Data was collected from multiple informants, including youth, caregiver, case manager, and teacher. We use data from Wave 3 (n = 5,872) given that this is the only wave in the NSCAW II that includes a measure of sexual orientation. The sample was restricted to those 11 years and older because questions about youth sexual orientation identity were not included in surveys of youth who were 10 and under. This provided a final sample of 1,309 youth.

Measures

Out-of-home placement

Two questions were used to reflect youth’s experience with out-of-home placements. First, we used the cumulative number of days that youth reported being in out-of-the home placements. Responses ranged from 0–1547 days. Second, we used the total number of out-of-home living arrangements. Responses ranged from 1–13 different placements. Those who reported no days of out-of-home placements or zero out-of-home living arrangements were coded as having no out-of-home placement = 0 and those with any reports of out-of-home placement, regardless of type, were coded as having experienced out-of-home placement = 1.

Sexual orientation

In Wave 3 of the NSCAW II, youth were asked to choose which of the following best fits how they think about themselves with response options of “totally straight (heterosexual),” “mostly straight but kind of attracted to people of your own sex,” “bisexual-that is attracted to males and females equally,” “mostly gay (homosexual) but kind of attracted to people of the opposite sex”, “totally gay (homosexual)”, and “not sexually attracted to either males or females”. Sexual orientation identity was recoded to reflect those that were totally straight or heterosexual = 0, those who reported any same-sex attraction or sexual minority identity (i.e., mostly straight, bisexual, mostly gay, and totally gay) were coded as sexual minority = 1 and youth who reported no sexual attraction to either males or females were coded as having no attractions = 2.

Analytic Strategy

First, after estimating basic descriptive statistics (see Table 4), we calculated a Disproportionality Representation Index (DRI; Morton, Ocasio, & Simmel, 2011; Wilson & Kastanis, 2015) to estimate whether sexual minority youth were overrepresented in child welfare and out-of-home care. More specifically, we divided the percentage of sexual minority youth in child welfare and out-of-home placement subgroups within the NSCAW II by estimates of sexual minority youth in the general population. This provides an index of disproportional representation of specific subgroups or populations within a sample. DRI values above 1.00 indicate overrepresentation and values under 1.00 indicate underrepresentation. To calculate the DRI, we used two general population data sources: Add Health, which was collected in 1994 (see Study 1), and the 2015 National Youth Risk Behavior Survey (Kann et al., 2016) which provided national estimate of sexual minority youth in middle and high schools. Second, we conducted chi-square test of independence to assess whether, among youth involved with Child Protective Services, sexual minority youth were more likely than heterosexual youth to report out-of-home placement. All analyses were design adjusted and weighted to reflect population estimates.

Table 4.

Study 2: NSCAW II Sample Characteristics (n = 1,014)

	n	%w
Biological Sex
Male	452	41.10
Female	562	58.90
Race/Ethnicity
White/Non-Hispanic	373	43.39
Black/Non-Hispanic	287	19.58
Hispanic	275	28.95
Other	62	5.54
Unknown	17	2.53
Parental Education
< High School	143	16.53
High School	302	26.80
Post High School	264	28.02
Missing	305	28.64
Out-of-home Placement
No	608	80.06
Yes	406	19.94
Sexual Orientation
Heterosexual	781	76.64
No attractions	86	7.63
Sexual minority (LGB)	147	15.73

Results

Youth over the age of 11 in Wave 3 of the NSCAW II were 52.33% female on the basis of sex (NSCAW does not include measures of gender identity), 34.99% White, 25.99% Black, and 24.37% Hispanic. The majority of youth reported that their parents had a high school diploma (29.79%), 25.06% with post high school education, and 14.59% with less than a high school diploma. Roughly, 41% of the sample indicated out-of-home placement and 14.35% of the overall sample were sexual minorities.

Next, we tested whether sexual minority youth were overrepresented in child welfare and the out-of-home placement subgroup by calculating DRIs using population estimates of sexual minority youth from Wave 1 of Add Health and the 2015 YRBS (see Table 5). Estimated DRIs indicate disproportionate rates of sexual minority youth in child welfare (DRIs ranging from 1.95 to 2.48) and out-of-home placement (DRIs = 3.69 to 4.68) when compared to what would be expected given rates of sexual minority youth in general adolescent populations.

Table 5.

Study 2: Disproportionality Representation Index of Sexual Minority Youth Involved in Child Welfare and Foster Care Placement Relative to the General Population

	General Population Metrics	LGB Youth in Child Welfare	LGB Youth in Foster Care
NSCAW % LGB		15.6%	29.5%
Add Health (1994)a	6.3%	DRIc = 2.48	DRI = 4.68
2015 Youth Risk Behavior Survey (2015)b	8.0%	DRI = 1.95	DRI = 3.69

^aages 12–18, based on measures of same-sex romantic attraction (see Study 1 of this paper)

^bages 12–18, based on measures of sexual identity (Kann et al., 2017)

^c$DRI=\frac{\%\mathrm{LGBQ}\mathrm{Youth}\mathrm{in}\mathrm{the}\mathrm{NSCAW}}{\%\mathrm{LGBQ}\mathrm{Youth}\mathrm{in}\mathrm{the}\mathrm{General}\mathrm{Population}}$

Results from chi-square analysis (Table 6) indicated that, among all youth involved in Child Protective Services, sexual minority youth were no more likely to report out-of-home placement than heterosexual youth, χ²(2, n = 1,014) = .47, p = .844. Similar to Study 1, we tested whether the prevalence of out-of-home placement varied across sexual orientation identity when lesbian/gay and bisexual youth were disaggregated: No statistical differences were detected, χ²(3, n = 1,014) = .18, p = .898.

Table 6.

Study 2: Chi-square Analyses Testing Adolescent Same-Sex Attraction among Child Welfare Involved Youth

	Full Sample
	No Out-of-Home Placement(s)			Had Out-of-Home Placement(s)			χ2,	p-value
	n	%w	(95% CI)	n	%w	(95% CI)
Sexual orientation
No attractions	53	83.01	(72.05, 90.25)	33	16.99	(9.75, 27.95)	0.47,	.844
Heterosexual	476	79.75	(75.07, 83.74)	305	20.25	(16.26, 24.93)
LGB	79	80.18	(71.38, 86.78)	68	19.82	(13.22, 28.62)

General Discussion

Using data from two nationally representative studies, the current study demonstrates an overrepresentation of sexual minority youth in child welfare, foster care, and out-of-home placement. In Study 1, we show that sexual minority youth are nearly 2.5 times more likely to be involved in the foster care system relative to heterosexual youth. Notably, post hoc analyses found that the effect was strongest among exclusively same-sex attracted youth. In Study 2 we found that sexual minority adolescents were disproportionately represented in a national sample of youth involved in child welfare and among adolescents who reported out-of-home placement (e.g., placement into foster or group homes). Contrary to expectations, once involved in the child welfare system, sexual minority youth were no more likely than heterosexual youth to report out-of-home placement, and thus appear to be no more or less likely than heterosexual youth to be removed from their home environments. This finding, coupled with the overrepresentation of sexual minority youth in out-of-home placement, may suggest that once in out-of-home placements, sexual minority youth may be less likely to exit child welfare custody compared to their heterosexual peers. This could be related to a lack of reunification efforts due to family rejection or lower chances of being adopted given that these youth would theoretically be adolescents when they recognize and assert a sexual minority identity (Dworski & Hall, 2013; Wilson & Kastanis, 2015). The lack of specificity in the available data related to the duration of out-of-home placement could also contribute to this null effect. Although there has not been a literature developed on how youths’ sexual minority status may influence the decisions-making ecology (Baumann, Fluke, Dlagleish, & Kern, 2013) of case workers and managers, there may be distinct decision-making processes unique to sexual minority youth that do or do not lead to out-of-home placement or reunification. Future research is needed in this area.

Our findings fill a critical gap outlined by the Administration for Children Families (Burwick et al., 2014): Namely, the need for nationally representative estimates of sexual minority youth involved in child welfare and foster care systems. Our results are wholly consistent with previous studies of community-based samples that demonstrate disproportionate rates of sexual minority youth in foster care and child welfare (Dworsky & Hall, 2013; Wilson & Kastanis, 2015). Such findings underscore the importance of future work in this area.

As attitudes towards sexual minority people have become more accepting in the United States, some may assume a diminishing need for concern (c.f. Russell & Fish, 2016; in press). Despite assumptions that sexual minority youth today are better off than in previous generations, a recent but robust literature demonstrates that today’s sexual minority youth continue to experience similar rates of disproportionate mental (Peter et al., 2017) and behavioral health burdens (Watson, Lewis, Fish, & Goodenow, 2018; Fish, Watson, Porta, Russell, & Saewyc, 2017) when compared to sexual minority youth from a decade ago. Relevant to sexual minority youth risk for child welfare entry, a recent study indicates that since 1998, sexual minority youth have not experienced the same increase in family connectedness and parental support (Watson, Saewyc, Adjei, & Rose, 2015) as heterosexual youth. In other words, sexual minority youth continue to experience disparities in parental support and connectedness and for some these disparities are getting wider (Fish & Russell, 2018; Watson et al., 2015). Considering narrative data of family rejection and pushout from homeless and child-welfare-involved sexual minority youth (Durso & Gates, 2012; Pearson et al., 2017), our findings suggest that contemporary sexual minority youth remain at elevated risk for child welfare entry. Of course, more work is needed to better understand mechanisms of parental rejection and pushout, and how these experiences relate to sexual minority youth becoming involved with child welfare, foster care, and out-of-home placement. Given our and other’s findings, it is imperative that research and resources be focused on developing a better understanding of the experiences of sexual minority youth once involved in the child welfare system.

Implications for Practice

Even with studies highlighting the systematic overrepresentation of sexual minority youth within the child welfare system, as well as unbalanced quality of care, there remains a lack of institutional guidelines and empirically supported practices to guide those working with sexual minority youth involved in child welfare services (Abbot et al., 2012; Burwick et al., 2014). For some sexual minority youth, for example, reunification may be uniquely dangerous and thus requires sexual minority informed services and considerations – such as mandated family counseling with a sexual minority affirming therapist – prior to approval (see Abbot et al., 2012). Such considerations also require that child welfare staff and social workers be trained in the unique service needs of sexual minority youth and the family dynamics that may result when youth come out in unsafe family and community environments (Ryan et al., 2009). Research on child protection is making progress in identifying contextual factors important for decision-making of placements (Fluke, Corwin, Hollinshead, & Maher, 2016; Graham, Dettlaff, Baumann, & Fluke, 2015). Several states require by law that out-of-home care is affirming, and that caregivers be educated about providing adequate care for sexual minority youth (Foster Care Services: Cultural Competency Bill, 2012; Human Rights Campaign, n.d.). Yet research suggests that the child welfare system is not well-prepared to provide safe and affirming care of sexual minority youth (Matarese, Greeno, & Betsinger, 2017; Wilber et al., 2006), and studies focused on the experience and needs of sexual minority youth in child welfare are woefully lacking. Continued focus on this population is needed to help inform policies and practices that improve the experiences of sexual minority youth involved with child welfare.

Limitations and Future Research

There are several limitations to note. First, the Add Health study asked participants about their romantic attractions, while the NSCAW II study asked about sexual orientation by combining terms used to describe sexual attraction with sexual identities (i.e., “attracted” and “gay”). The lack of consistent and comprehensive measures of sexual orientation in national surveys makes it difficult to infer the different meanings of these constructs for youth, but also to make detailed comparisons across studies (Temkin, Belford, McDaniel, Stratford, & Parris, 2018). Further, with the current data we cannot conclude whether youth disclosed their romantic or sexual attractions to others, and more importantly to their family of origin. Ideally, studies that include multiple measures of sexual minority status and the disclosure of this status would better contextualize the experiences of sexual minority youths’ overrepresentation in child welfare. Second, although the child welfare sample from NSCAW II is recent (surveyed in 2013), Add Health is dated (surveyed 1994–2002). Such limitations require that population-based data collection efforts include measures of sexual minority status as well as living arrangements in order for better estimates of contemporary rates of sexual minority youth overrepresentation in child welfare services. We also were unable to explore differences on the basis of gender identity given that these surveys did not include these measures. Third, given the number of youth who report same-sex attraction and foster care involvement, we were unable to test differences on the basis of sex. Such investigations could document important differences in risk for foster care entry among sexual minority youth. We did conduct post hoc tests to assess if there were differences in the prevalence of foster care (study 1) and out-of-home (study 2) placement for exclusively same-sex attracted youth and bisexually-attracted youth, although analyses were likely underpowered and should be interpreted with caution. Fourth, because the surveys used in the current study were not designed to understand the distinct experiences of sexual minority youth, we have no information about minority-specific mechanisms that may contribute to overrepresentation in foster care, such as parental rejection, discrimination, and stigma. Combining data from large representative studies with quantitative and qualitative data from sexual minority youth and their caregivers may provide more in-depth information about the mechanisms that explain the overrepresentation of sexual minority youth in foster care. Such research approaches would likely yield important recommendations for child welfare services.

Table 3.

Study 1: Adjusted Logistic Regression Models of Sexual Attraction Predicting Foster Care Involvement in Add Health (N =14,154)

	aOR	(95% CI)	p-value
Wave 1 Attraction
Different-sex attraction[ref]	1.00
No attractions	1.37	(0.86, 2.17)	.187
Any same sex attraction	2.43	(1.40, 4.21)	.002

Note. Controlling for sex, race/ethnicity, parental education, receipt of public assistance. Multiple imputation was used to account for missing data on covariates.