Understanding organizations for runaway and

3 downloads 0 Views 969KB Size Report
Robert Freeman c, Elizabeth Silverman a, Alexandra Kutnick a, Victoria ... We found settings overall evidenced satisfactory-to-high quality on a multi- ... and provides guidance on strategies to assess setting quality. .... Miner, Walker, & Davidson, 2007). ..... 6) trading sex for money, drugs, or a place to stay; and 7) pimping.
Children and Youth Services Review 73 (2017) 398–410

Contents lists available at ScienceDirect

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

Understanding organizations for runaway and homeless youth: A multi-setting quantitative study of their characteristics and effects Marya Viorst Gwadz a,⁎, Charles M. Cleland a, Noelle R. Leonard a, James Bolas b, Amanda S. Ritchie a, Lara Tabac c, Robert Freeman c, Elizabeth Silverman a, Alexandra Kutnick a, Victoria Vaughan Dickson d, Margo Hirsh c, Jamie Powlovich b a

Center for Drug Use and HIV Research, Rory Meyers College of Nursing, New York University, New York, NY, United States Coalition for Homeless Youth, New York, NY, United States c Independent Consultant, New York, NY, United States d Rory Meyers College of Nursing, New York University, New York, NY, United States b

a r t i c l e

i n f o

Article history: Received 13 October 2016 Received in revised form 16 January 2017 Accepted 17 January 2017 Available online 18 January 2017 Keywords: Runaway youth Homeless youth Program quality Services Positive Youth Development Youth Program Quality Assessment

a b s t r a c t Runaway and homeless youth (RHY) are served by specialized settings (e.g., Drop-In Centers, Transitional Living Programs, and multi-program settings), but little is known about the characteristics of these organizations or their effects on RHY’s behavioral and psychosocial outcomes. To address this gap we studied 29 randomly selected diverse settings across New York State, including those in rural, suburban, and urban areas. Within settings, we used the Youth Program Quality Assessment model to observe and rate programs for RHY (N=53), assess program administrators (N = 30), and conduct anonymous structured assessment batteries with RHY aged 16-21 years (N = 463). We found settings overall evidenced satisfactory-to-high quality on a multiperspective setting quality score. With respect to RHY’s behavioral outcomes, engagement in school/job training/work was high (81 %), substance use was moderate (mean use: 17 of the past 90 days), and 37 % evidenced involvement in the street economy (e.g., drug dealing, burglary). RHY in Transitional Living Programs and multi-program settings had more engagement in school/job training/work and less involvement in the street economy than their peers in Drop-In Centers. The quality of settings was not associated with these three behavioral outcomes (school/training/work, substance use, street economy), likely due to issues of restricted range. However, higher setting quality was associated with four constructive psychosocial outcomes; namely, RHY’s perceptions that settings foster positive outcomes in these three domains, and perceived resilience. Thus the present study highlights settings’ overall good quality, with some variability, and provides guidance on strategies to assess setting quality. Consistent with the existing literature, RHY in Drop-In Centers are highly vulnerable and may require additional types of services/programs to achieve their potential. Further, while the present study suggests all settings benefit RHY, better quality settings may be able to move beyond meeting RHY’s basic requirements and address higher order relational, psychosocial, and motivational needs. Importantly, fostering a sense of resilience among RHY, as well as young people’s experiences of settings as helpful to them in achieving good behavioral outcomes, may have long-term beneficial effects on RHY’s engagement in other settings, relationships, adaptation, and functioning. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction Runaway and homeless youth (RHY) are young people between 13 and 24 years of age who have run away from or been forced to leave their homes, who reside without parental/guardian supervision in temporary situations, places not intended for habitation, or emergency shelters (Whitbeck, Hoyt, & Bao, 2000). Although precise figures are lacking, ⁎ Corresponding author at: Center for Drug Use and HIV Research, Rory Meyers College of Nursing, New York University, 433 First Avenue, New York, NY 10010, United States. E-mail address: [email protected] (M.V. Gwadz).

http://dx.doi.org/10.1016/j.childyouth.2017.01.016 0190-7409/© 2017 Elsevier Ltd. All rights reserved.

as many as 2.8 million youth are estimated to be homeless in the U.S. each year (Cooper, 2006; National Alliance to End Homelessness, 2012). Many of these young people return home within a week, but a substantial proportion remain out-of-home for substantial periods of time, or even permanently (Tevendale, Comulada, & Lightfoot, 2011). It is well documented that RHY experience high rates of physical, emotional, and sexual abuse, as well as neglect, other traumas, and chronic stress throughout their lifespans. However, they have only minimal involvement in the systems and settings that typically foster and protect young people, such as supportive families, pro-social peers, safe communities, and schools (Bao, Whitbeck, & Hoyt, 2000; Gwadz

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

et al., 2009; Gwadz, Nish, Leonard, & Strauss, 2007). As a result, RHY evidence high rates of serious relational, mental health, physical health, psychosocial, and behavioral problems. These problems, in turn, place them at grave risk for adverse long-term outcomes, including chronic unemployment, entrenchment in the street economy (e.g., drug dealing, transactional sex/being trafficked), hazardous substance use, incarceration, adult homelessness, unstable relationships, poor health, and even early mortality (Cleverley & Kidd, 2011; Gwadz et al., 2010; Tucker, Edelen, Ellickson, & Klein, 2011). African American/Black and Hispanic young people, those with lesbian, gay, bisexual, and other non-heterosexual sexual orientations, and individuals with transgender gender identities are over-represented among RHY compared to the general population (Cochran, Stewart, Ginzler, & Cauce, 2002; Keuroghlian, Shtasel, & Bassuk, 2014). These minority group statuses influence youths' patterns of risk, in part due to aspects of the larger environment, such as youth from these minority groups facing a greater likelihood of being stopped by police among compared to their White, heterosexual, and/or gender normative peers (Himmelstein & Brückner, 2011; Snapp, Hoenig, Fields, & Russell, 2015). Additionally, these minority group statuses affect RHY's treatment needs (Keuroghlian et al., 2014). Yet in the context of these serious risk factors, RHY evidence resilience—the capacity to withstand or recover from significant challenges that threaten an individual's stability, viability, or development (Masten, 2011). For example, leaving home is a type of coping response, and surviving out-of-home requires resourcefulness and adaptability (Bender, Thompson, McManus, Lantry, & Flynn, 2007; Rew, Taylor-Seehafer, Thomas, & Yockey, 2001). Resilience can be fostered among those suffering adversity through structured interventions and close, supportive relationships (Ungar, 2013; Zolkoski & Bullock, 2012). Moreover, experiencing or perceiving oneself as resilient is a potent correlate of persistence, resourcefulness, self-efficacy, and resultant positive behavioral and mental health outcomes (Sapienza & Masten, 2011). Among RHY, perceived resilience is associated with less suicidal ideation (Cleverley & Kidd, 2011) and fewer life threatening behaviors, such as attempted suicide (Rew et al., 2001). Perceived resilience may be vital for RHY because they lack the social and organizational resources available to typically developing adolescents.

1.1. Settings for RHY Across the U.S., a network of specialized programs has emerged to locate, engage, house, support, and treat RHY. These include the Basic Center Program to provide short-term programs (30 days or less) for RHY under 18 years of age, as well as long-term programs, including Transitional Living Programs (TLP) and Drop-in Centers (DIC) (New York State Office of Children & Family Services, 2014). TLPs are supported residences where RHY can reside for up to 18 months. TLPs typically provide counseling in basic life skills, interpersonal skills building, educational advancement, job attainment skills, and physical and mental health care (Family and Youth Services Bureau, 2016). DICs tend to take a “low-threshold” approach, providing a safe and supportive space that is easy to access. In DICs, RHY can socialize and rest, and receive tangible services (food, laundry, showers), mental health counseling, health services, and street outreach. DICs are appropriate for RHY who are not ready to or uninterested in entering residential or higher threshold programs, or who are waiting for a residential placement to become available. Importantly, DICs seek to engage street-based RHY who do not present for services elsewhere. Both TLPs and DICs foster the ultimate goal of preparing RHY for successful future independent living (The National Network for Youth, 2015). Yet RHY providers report funding levels for RHY programs are generally insufficient to meet need, particularly regarding housing for RHY, and further, funding levels tend not to be either stable or predictable over time (The National Network for Youth, 2013; United States Interagency Council on Homelessness,

399

2016). These types of fiscal factors and constraints complicate settings' efforts to meet the needs of this vulnerable population. In light of the grave challenges to psychosocial development RHY face throughout their lives and the difficulties inherent in engaging and treating them (Slesnick, Meyers, Meade, & Segelken, 2000), the importance of these specialized settings cannot be overstated (Thompson, Bender, Windsor, Cook, & Williams, 2010). Yet the empirical literature is scant on settings for RHY, and on settings' effects on RHY's behavioral and psychosocial functioning (Karabanow & Clement, 2004). A body of work exists on specific behavioral interventions conducted in RHY settings (Altena, Brilleslijper-Kater, & Wolf, 2010; Slesnick et al., 2016), and some research has described individual programs, or a small set of programs (Altena et al., 2010; Heinze, Jozefowicz, & Toro, 2010; Pollio, Thompson, Tobias, Reid, & Spitznagel, 2006; Woods, Samples, Melchiono, & Harris, 2003). However, almost no research to date has sought to understand the characteristics of settings that serve RHY more broadly, or how settings may influence RHY's behavioral and psychosocial functioning. Yet settings for RHY are commonly called upon to demonstrate the efficacy of the specialized services they provide (Kidd, Miner, Walker, & Davidson, 2007). The present study addresses these gaps in the literature. 1.2. Conceptual model RHY programs are authorized by the Runaway and Homeless Youth Act (Juvenile Justice and Delinquency Prevention Act, Pub. L. 93-415, Sept. 7, 1974, 88 Stat.1109 (Title 42, Sec. 5601 et seq.) (Levesque, 2011; Slesnick, Dashora, Letcher, Erdem, & Serovich, 2009). Consistent with guidance in this Act, most RHY settings are guided by the Positive Youth Development approach (Levesque, 2011), a strengths-based model for encouraging resilience and self-sufficiency among youth, emphasizing the importance of youths' engagement in their own development and goals (Eccles & Appleton-Gootman, 2002). The Youth Program Quality Assessment (YPQA) model (Wilson-Ahlstrom, Yohalem, DuBois, & Ji, 2011; Yohalem, Wilson-Ahlstrom, Fischer, & Shinn, 2009) has emerged from the Positive Youth Development approach, and it provides a framework for conceptualizing and assessing the quality of out-of-school settings (e.g., after-school programs) that promote constructive development among youth. The YPQA model (Fig. 1) frames program quality into two broad categories: offeringlevel and organizational-level characteristics. Offering-level characteristics refer to the social processes and interactions that youth experience when they are engaged in the setting (e.g., a sense of safety or belonging). Organizational-level characteristics include expectations, policies, practices, and accessibility that support the production of high-quality youth experiences (e.g., whether policies and practices are youthcentered). The two domains correspond to the structure of the typical youth-serving organization: offerings within an organization. 1.3. The current study This study has two main aims. First, we describe the overall quality of settings for RHY, grounded in the YPQA model. Second, we explore whether RHY in higher quality settings evidence superior functioning in a number of key behavioral and psychosocial domains, to better understand both whether RHY settings foster positive outcomes among youth and whether setting quality plays a role in RHY's outcomes. The present cross-sectional study focuses on long-term settings for RHY in a large, discrete geographical area, New York State, and on RHY aged 16 to 21 years old, referred to as “youth” and “young people” in the present paper. As noted above (Sections 1 and 1.1), RHY tend to develop along an atypical path, and settings vary in size, structure, geographical context, and programs offered. The present study examines three behavioral outcomes critical for RHY's positive development that are also typically prioritized in long-term RHY settings: involvement in school, training,

400

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

Fig. 1. Youth Program Quality Assessment (YPQA) model.

and/or work; reduced frequency of substance use; and prevention of involvement in the street economy (e.g., drug dealing, being trafficked/ transactional sex, burglary). We also examine psychosocial outcomes; namely, RHY's perspectives on whether settings help them in these domains, as well as perceived resilience. We speculate RHY's perceptions that settings are helpful may have long-term beneficial effects on their engagement in other settings and relationships with professionals, and therefore, on their adaptation, and functioning. Moreover, perceived resilience is critical for the population of RHY, as described above. We hypothesized that RHY in settings of higher quality would have more favorable behavioral outcomes in these domains, be more likely to report the setting helps them achieve positive outcomes, and have greater perceived resilience. This is because, as suggested by the YPQA model, higher setting quality allows organizations to not only meet RHY's basic needs, but also move beyond these to provide higher order program offerings that engage, motivate, and build relationships with RHY, as shown in Fig. 1. 2. Methods 2.1. Overview of design and sample The YPQA guided a multi-level quantitative assessment of settings serving RHY that involved observations of specific programs offered within settings, which were coded on a quantitative rating scale, and assessments of program administrators' and RHY's perspectives on settings. Settings were defined as organizations providing one or more long-term programs for RHY. Settings were located in New York State, a region with N50 organizations serving RHY, found “downstate,” namely, the densely urban New York City metropolitan area, and “upstate,” which comprises urban, suburban, and rural areas. A stratified, multistage random sampling approach captured diversity in setting types and their geographical locations (suburban, rural, and urban). In

cases where settings closed or changed program type during the study period (N = 5 settings), they were replaced with a randomly selected setting of the same type and location. From a total of 50 settings, 29 settings were randomly selected for inclusion. Study procedures were approved by Institutional Review Boards at New York University and SolutionsIRB. 2.2. Sources of data The YPQA is a validated instrument designed to assess specific offering- and organizational-level characteristics of non-academic settings for youth (Smith & Hohmann, 2005; Yohalem et al., 2009). Although not developed specifically for RHY settings, the YPQA has previously been used in, and is considered appropriate for, RHY settings (R. Ramaswamy, personal communication, January 10, 2015). Form A of the YPQA uses direct observation of program activities to assess offering-level characteristics, namely safe environment, supportive environment, interaction, and engagement (Fig. 1). Form B uses interviews with program administrators to assess organizational-level characteristics, namely high expectations for youth and staff, youthcentered policies and practices, and access. The quality of youth experiences in settings with multiple program offerings is captured by aggregating ratings to the setting level (Smith & Hohmann, 2005). Training on administration of the YPQA was provided by the David P. Weikart Center for Youth Program Quality (http://www.cypq.org) using a video reliability check method. In addition to YPQA Forms A and B, we conducted an anonymous structured assessment battery with RHY. Fig. 2 depicts the multi-level assessment strategy within settings. 2.3. Procedures Experienced researcher team members from disciplines including anthropology, sociology, social work, and psychology made an on-site

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

401

Fig. 2. Multi-level data sources.

visit (usually 2–3 team members, 2–3 days per setting) to each setting in 2014–15. Prior to the site visit and in consultation with the setting's executive director, we identified the program administrator, defined as the individual with primary responsibility for administering the setting's RHY program(s). The program administrator served as the primary liaison to the research team and guided study activities, as described below.

2.3.1. Assent/consent Program administrators, all of whom were aged 18 years or older, and RHY aged 18 years and older gave verbal informed consent prior to participation, including in observations. A waiver of parental consent was obtained from the Institutional Review Board for RHY ages 16 or 17 years. These young people provided informed assent, and in addition, the approval of a Child Advocate (a staff member in the setting with knowledge of and a relationship with the young person) was elicited before the youth could participate in activities.

2.3.2. Eligibility criteria Eligibility criteria for RHY were 1) age 16–21 years; 2) active client of the setting (i.e., has completed an intake at least one month ago and attended a program at the setting at least twice in the past month or resided there for at least two weeks); 3) if aged 16 or 17 years, approval of a Child Advocate; and 4) not currently enrolled in foster care.

2.3.3. Compensation Each setting received compensation for the time and effort needed to host the research team ($1000). Program administrators were compensated $150 for participating in confidential interviews. Youth were compensated $25 for their participation in the anonymous assessment battery.

2.4. Data collection procedures 2.4.1. Observations (YPQA Form A) Program Administrators guided the selection of representative observable program offerings (e.g., a support group, small group discussion, casual grouping of RHY in a DIC, or cooking class). Following standard procedures for YPQA administration, two trained research staff members simultaneously observed two representative program offerings, where possible. Observations were coded independently in an online numerical data entry form. 2.4.2. Coded interviews with Program Administrator (YPQA Form B) The Program Administrator was interviewed individually by a single trained research staff member using a semi-structured guide. Responses were coded by the research staff member on an online numerical data entry form. 2.4.3. Anonymous structured assessments with RHY In small settings, all RHY were approached and recruited, and in larger settings, a systematic approach was used to reduce selection bias and provide RHY with equitable access to the assessment, such as recruiting the first 25 clients who presented at the program on a certain day. Assessments lasted 45–60 min and were conducted using an audio computer-assisted self-interview (ACASI) program. 2.5. Instruments 2.5.1. YPQA (Form A/observation and Form B/interview) For both the offering-level characteristics assessed with Form A and the organizational-level characteristics assessed with Form B, individual YPQA items were scored as 0, 2, or 4, with higher scores indicating better quality. Scale scores were created by averaging item scores. Internal consistency reliability (coefficient alpha) was 0.62, 0.88, 0.90, 0.80, 0.66, 0.81, and 0.55 for safe environment, supportive environment,

402

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

interaction, engagement, high expectations for youth and staff, youthcentered policies and practices, and access scales, respectively. 2.6. Assessment battery for RHY 2.6.1. Socio-demographic, background, and health-related factors Sex at birth, age in years, race, ethnicity, transgender gender identity, sexual orientation, education, involvement with the criminal justice system, history of homelessness, foster care/group home placement history, and the types of services used at the setting were measured using the RHY Status Scale (Gwadz et al., 2010). We assessed symptoms of depression over the past week using the eight-item Center for Epidemiologic Studies Depression Scale (CES-D8; α = 0.81) (Turvey, Wallace, & Herzog, 1999), and current anxiety using the Brief Symptom Inventory (BSI) anxiety subscale (6 items; α = 0.89) (Derogatis & Melisaratos, 1983). We calculated composite depression and anxiety scores using cutoffs of 7 or greater and 0.7 or greater, respectively, to determine presence or absence of symptoms at a clinically significant level (Derogatis & Melisaratos, 1983; Radloff, 1977). Lifetime substance use (drug and/or alcohol use) was assessed based on affirmative responses to one or more of 8 items, “Did you ever use” alcohol, marijuana, crack, cocaine, heroin, nonprescription methadone, painkillers/opiates, other drug (Graves, 1995). 2.6.2. Program quality The Supports and Opportunities Scale (SOS) (Youth Development Strategies Inc, n.d., unpublished measure) assessed RHY's perceptions

of setting quality across four domains that correspond to the YPQA. The supportive relationships subscale (13 items) assessed the extent to which the setting provides guidance, emotional support, and practical support (α = 0.91). The safety subscale (9 items) measured physical and emotional safety (α = 0.86). Youth involvement (16 items) assessed how well the setting promoted decision-making, youth leadership, and a sense of belonging (α = 0.91). Skill building (7 items) measured the degree to which the setting and its offerings were experienced as interesting, challenging, and promoting growth (α = 0.88). Items were rated on a 5-point Likert-type scale from 0 (strongly disagree) to 4 (strongly agree). 2.6.3. Overall setting quality A score for overall setting quality was created by averaging the following eleven scores: 1) SOS supportive relationships; 2) SOS safety; 3) SOS youth involvement; 4) SOS skill building; 5) Form A safe environment; 6) Form A supportive environment; 7) Form A interaction; 8) Form A engagement; 9) Form B youth-centered policies and practices; 10) Form B high expectations for youth and staff; and 11) Form B access. Overall setting quality score could range from 0 to 4, with higher values indicating better setting quality. Coefficient alpha for overall setting quality was 0.78. 2.6.4. Behavioral and psychosocial outcomes Youth were considered engaged in school, work, or job training if they answered “Yes” to any of the following (Gwadz et al., 2010): 1) “Have you gone to school in the past 3 months?” 2) “Have you

Table 1 Description of settings (%).

Description of settings (%) DIC (N = 9)

TLP (N = 11)

Dual or multiprogram (N = 9)

Total (N = 29 settings)

Geographical location Rural

11.11

45.45

11.11

24.14

Suburban

11.11

18.18

22.22

17.24

Urban

77.78

36.36

66.67

58.62

Upstate

55.56

63.64

55.56

58.62

Downstate

44.44

36.36

44.44

41.38

Housing–TLP

13.67

86.01

51.03

52.71

Housing–crisis shelter

21.31

9.38

17.74

15.68

Case management

56.45

73.99

63.60

65.32

Showers

36.60

94.17

57.08

64.79

Laundry

38.10

85.45

47.08

58.84

Food pantry

42.96

75.70

44.41

55.83

Type of location

Services received by surveyed youth

Outreach services

48.73

20.37

41.55

35.74

Health education

29.33

53.70

37.14

41.00

Support groups

40.60

32.86

32.65

35.20

Job training

32.51

31.25

44.43

35.74

Health care

18.99

39.50

19.89

27.05

Mental health counseling

25.34

33.33

25.16

28.32

Job placement assistance

13.55

24.01

35.46

24.33

Substance use counseling

8.15

7.95

13.04

9.59

Note: Shaded rows indicate differences among groups at a statistically significant level.

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

403

Table 2 Setting quality (Mean, SD).

Setting level (range 1.81–3.70) Form A – Observations Safe environment Supportive environment Interaction Engagement Form B – Program Administrators Youth Centered Policies and practices High expectations for youth and staff Access RHY level SOS (range 1.80–3.14) Safety Supportive relationships Skill building Youth Involvement Total setting quality score (0–4)

DIC (N = 9 settings)

TLP (N = 11 settings)

Dual or multi-program (N = 9 settings)

Total (N = 29 settings)

3.46 (0.38) 2.54 (1.38) 1.91 (1.19) 1.81 (1.13)

3.70 (0.28) 3.23 (1.17) 3.00 (1.34) 2.28 (1.71)

3.46 (0.38) 3.67 (0.30) 3.07 (0.57) 2.80 (1.06)

3.55 (0.36) 3.11 (1.16) 2.64 (1.22) 2.32 (1.41)

2.37 (0.63) 2.84 (0.92) 3.48 (0.39)

2.39 (0.54) 3.56 (0.50) 3.22 (0.75)

2.83 (0.78 3.40 (0.73) 3.36 (0.86)

2.53 (0.66) 3.30 (0.75) 3.34 (0.69)

3.06 (0.78) 2.83 (0.85) 2.66 (0.89) 2.07 (0.79) 2.68 (0.50)

3.14 (0.75) 2.95 (0.87) 2.66 (0.89) 1.80 (0.62) 2.94 (0.42)

2.93 (0.72) 2.81 (0.74) 2.59 (0.83) 1.92 (0.76) 3.02 (0.29)

3.00 (0.75) 2.83 (0.80) 2.62 (0.86) 1.95 (0.76) 2.89 (0.43)

gone to a General Educational Development (GED) program in the past 3 months?” 3) “Are you currently going to a regular school, not including a GED program?” 4) “Are you currently going to a GED program?” 5) “Are you currently going to a job training program or another form of education or job skills program?” 6) “Have you gone to a job training program or another form of education or job skills program in the past 3 months?” 7) “In the past 3 months, have you held a job where you were paid with a check?” and 8) “In the past 3 months, have you held a job that was off-the-books?” Using a subscale of the Global Appraisal of Individual Needs–Initial (GAIN-I) instrument (Dennis, Titus, White, Unsicker, & Hodgkins, 2002), recent substance use was assessed by asking RHY the number of days in the past 90 days on which each of the following substances were used, using “street” names for drugs as appropriate: 1) alcohol; 2) marijuana; 3) crack cocaine; 4) other forms of cocaine; 5) heroin; 6) nonprescription/street methadone; 7) painkillers, opiates, analgesics; 8) hallucinogens; 9) anti-anxiety drugs or tranquilizers; 10) methamphetamine; 11) amphetamines/stimulants; 12) barbiturates or sedatives; and 13) any other drug. We coded the number of days using one of more substances in the past 90 days as the maximum number of days across these types. For analysis, number of days was transformed to proportion of days, and then the proportion was transformed to a logit. Recent involvement in the street economy (yes/no) was assessed by asking youth whether they engaged in any of the following activities in the past three months: 1) panhandling; 2) shoplifting or other theft; 3) breaking into a house, store, or car; 4) robbery; 5) drug dealing; 6) trading sex for money, drugs, or a place to stay; and 7) pimping someone (Gwadz et al., 2010). Using a measure of program involvement (Gwadz et al., 2010), youth were asked how much the programs they attend and/or services they receive 1a) “help you achieve your goal of going back to school”; 1b) “help you stay in your educational program”; 1c) “help you keep a job where you are paid with a check”; 1d) “help you to get a job where you are paid with a check”; 2a) “help you manage your drug or alcohol use”; 2b) “help you avoid drug and alcohol use”; 3) “help you to avoid working in the street economy” (if so desired). Possible responses to these interview items were not at all; a little; somewhat; quite a bit; and very much. Youth who responded with quite a bit or very much were coded as helped by the setting. Perceived resilience was assessed with the 10-item Connor– Davidson Resilience Scale (CD-RISC), a unidimensional scale with good internal consistency (coefficient alpha = 0.90 in this sample) and construct validity (Campbell-Sills & Stein, 2007). Items such as “I believe I can achieve my goals, even if there are obstacles” were rated on a 5-point Likert-type scale from 0 (Not at all true) to 4 (True nearly

all of the time). The ten individual items were summed, yielding a score that could range from 0 to 40. 2.7. Statistical analysis Across the 29 settings, we conducted 53 observations at 23 settings; 30 program administrator interviews at 29 settings; and assessment batteries with 463 RHY at 29 settings. Mixed-effects regression models (Bates, Maechler, Bolker, & Walker, 2015; Kuznetsova, 2015) with random intercepts to account for nesting of youth in settings were used to examine associations between settinglevel characteristics (type, location, and overall quality rating) and youth-level outcomes. Youth-level covariates included the following: gender; transgender; age; sexual orientation; time out of home; time engaged in drop-in program; and time engaged in transitional living program. Continuous variables were centered and categorical variables were effects coded such that the intercept represented the expected outcome for an average youth in an average setting. Given the small number of settings, coefficients for setting type, location, and overall quality rating were estimated in separate models. All tests of statistical significance were two-tailed, and p b 0.05 was considered significant. The R statistical computing environment (R Core Team, 2015) was used for all analyses. 3. Results Table 1 shows settings' location and geographical distribution, organized by type of setting. DIC and dual/multi-program settings were less likely to be located in rural and suburban locations than in urban locations, and TLP settings were more likely to be in rural locations than in suburban or urban locations. There were more settings of every type located upstate than downstate. Youth in TLP settings were more likely to be living in a TLP, not surprisingly, and to receive shower and laundry services than RHY in the other types of settings. Youth in dual-multi program settings with two or more program components were more likely to receive job training and placement assistance than RHY in settings with only one component i.e., TLP or DIC). (In Tables 1–5, tests of significance not shown for parsimony but indices with statistically significant differences [p b 0.05] are highlightEd.) Table 2 summarizes ratings of setting by observers (Form A), program administrators (Form B), and youth (SOS). There were no statistically significant differences in setting quality ratings by setting type. Fig. 3 shows overall setting quality ranked and indicates setting type and the number of youth surveyed for all 29 settings. The multi-perspective setting quality scale ranges from 0 to 4, with higher values indicating higher quality. Only setting quality values from 2.0 to 3.5 were observed among the 29 settings studied. Thus all settings fell either at or above

404

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

the middle of the scale (a score of 2.0 or higher), with 44.8% of settings scoring 3.0 or higher. Table 3 presents characteristics of surveyed youth, including demographics, out-of-home history, and living situation in the past three months. Youth in TLP settings were more likely to be nonHispanic White, female, and heterosexual than RHY in other types of settings. Youth in TLP settings also were more likely to have had stable housing in the past three months (yet not all RHY in TLPs resided only at the TLP in the past 3 months), and were less likely to have had severe economic need in the past month than RHY in other types of settings.

Youths' risk and protective factors are shown in Table 4. Youth in TLP settings were less likely to have been in jail and less likely to have been involved in the street economy recently and lifetime than RHY in the other types of settings. Table 5 describes youth outcome variables. Most youth (81%) were currently engaged in school/job training/work, although this was somewhat less likely among RHY in DIC settings. More than half of RHY surveyed reported settings helped them to remain engaged in school, training, or work. On average, substance use occurred on about 17 of the past 90 days. Almost two-thirds of RHY (64%) reported substance use in the past 90 days, but only 17% reported use of substances other

Table 3 RHY's socio-demographic and background factors (Mean [SD] or %).

RHY’s socio-demographic and background factors(Mean [SD] or %) DIC (N = 156)

TLP (N = 61)

Dual or multiprogram (N = 246)

Total (N = 463)

19.81 (1.30) 42.31 34.62 18.59 49.36

18.23 (1.53) 34.43 24.59 54.10 73.77

19.26 (1.48) 51.22 34.96 16.26 54.47

19.31 (1.51) 46.00 33.48 22.03 55.29

Transgender gender identity –female-to-male Male sex at birth Transgender gender identity –male-to-female

19.48 50.64 11.39

2.22 26.23 6.25

14.18 45.53 4.46

13.67 44.71 7.25

Non -heterosexual sexual orientation (LGBQ) History of being out-of-home Age first out-of-home (runaway, thrown away, kicked out) Ever in foster care or group home Number of foster care placements (including group home) Total number of years in foster care/group home Been out of home >3 years Currently considers self to be homeless Where lived in past three months Temporary crisis shelter

55.13

26.67

40.74

43.79

16.01 (7.02)

15.21 (3.91)

15.77 (3.84)

15.78 (5.24)

44.23 8.52 (17.88)

40.98 3.20 (2.06)

33.74 6.76 (14.68)

38.23 6.94 (15.08)

5.80 (6.67) 44.29 56.77

4.60 (4.92) 38.60 47.46

4.28 (5.07) 34.06 55.51

4.92 (5.74) 38.03 54.90

25.33

10.00

34.31

28.06

73.33 8.33 20.00 1.67 20.00 26.23 58.33

32.64 20.50 31.38 7.11 24.27 28.86 37.66

32.52 20.71 29.18 6.46 26.28 26.35 38.98

48.21 18.03

46.89 35.95

47.51 33.55

41.67 49.02

49.80 52.36

46.64 49.09

68.85 47.54 9.84 18.03

54.07 40.65 14.63 43.09

52.48 36.07 15.12 35.64

Age in years African American/Black Latino/Hispanic White/other Female sex at birth

TLP 16.00 In own home or rented room 26.00 In someone else’s home 29.33 In a place not intended for habitation 7.33 Other 32.00 Lived in >1 place in past 3 months 22.44 Stably housed over the past 3 months (i.e., lived in 33.33 only 1 appropriate and potentially long-term location) Earned less than $100 in the past month 48.28 Severe economic need (did not have food for 2 days 35.95 or more in past month) Education, job training, employment High school diploma or equivalent 43.59 High school diploma or equivalent (among those age 44.16 17+years) Currently in school or GED program 43.59 Currently has job that is on-the-books 24.36 Currently has job that is off-the-books 17.95 Currently in job training program 30.77 Note: Shaded rows indicate differences among groups at a statistically significant level.

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

405

Table 4 Risk and protective factors among RHY (Mean [SD] or %).

Risk and protective factors among RHY (Mean [SD] or %) DIC (N = 156)

TLP (N = 61)

Number of days enrolled in DIC 275.08 (287.01) 5.66 (23.15) Number of days enrolled in TLP 19.81 (93.38) 148.54 (169.68) Physical and mental health Health very good or excellent 59.62 60.66 Depression at clinically significant 59.09 62.30 level Anxiety at clinically significant level 69.23 80.33 Involvement in the criminal justice system Ever picked up by police 69.23 63.93 Ever been in jail 54.49 29.51 Age first in police custody, jail, prison 17.06 (8.53) 15.33 (1.82) (years) Total amount of time ever spent in 15.38 6.56 detention, jail, or prison is greater than six months Setting helps youth avoid being 55.13 68.85 picked up by the police or being put in jail or prison Substance use Ever smoked cigarettes 67.95 75.41 Ever used alcohol 69.23 81.97 Ever used marijuana 65.38 75.41 Ever took painkillers/opioids/other 18.59 21.31 analgesics Ever used other drugs (crack, heroin, 24.36 27.87 etc.) Ever injected drugs 5.77 3.28 Ever had a problem with alcohol 16.45 22.41 Ever had a problem with marijuana 11.11 20.34 Ever had a problem with other drugs 9.62 9.84 Involvement in street economy – lifetime Panhandling 42.31 21.31 Drug dealing 38.46 22.95 Being trafficked/transactional sex 28.21 11.48 (traded sex for money, drugs, etc.) Theft, shoplifting 53.85 47.54 Mugged or robbed someone 22.44 3.28 Breaking into house, store, or car 17.31 13.11 Pimping someone 12.82 4.92 Involvement in street economy – past 3 months Panhandling 23.72 1.64 Drug dealing 16.67 0.00 Being trafficked/transactional sex 7.69 6.56 (traded sex for money, drugs, etc.) Theft, shoplifting 23.08 6.56 Mugged or robbed someone 7.69 0.00 Breaking into house, store, or car 2.56 1.64 Pimping someone 3.21 0.00 Note: Shaded rows indicate differences among groups at a statistically significant level.

Dual or multiprogram (N = 246) 102.07 (212.69) 72.03 (150.20)

147.66 (246.87) 64.52 (142.36)

59.35 63.52

59.61 61.87

75.61

74.08

66.67 47.15 15.73 (2.58)

67.17 47.30 16.22 (5.06)

8.54

10.58

62.20

60.69

61.79 75.61 68.70 19.11

65.66 74.30 68.47 19.22

17.48

21.17

1.63 13.58 16.80 5.28

3.24 15.67 15.35 7.34

23.58 32.93 23.17

29.59 33.48 23.33

53.66 11.79 18.70 10.16

52.92 14.25 17.49 10.37

13.01 13.41 6.91

15.12 12.74 7.13

17.07 2.85 3.25 2.44

17.71 4.10 2.81 2.38

Total (N = 463)

406

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

Table 5 RHY's behavioral and psychosocial outcomes, by type of setting (Mean [SD] or %).

RHY’s behavioral and psychosocial outcomes, by type of setting (Mean [SD] or %) DIC (N = 156)

TLP (N = 61)

Dual or multiprogram (N = 246)

Total (N = 463)

Currently in school, job training, and/or employed in the formal economy

67.95

91.80

86.59

80.99

Setting helps with school, job training, job

52.26

75.41

64.34

61.74

19.35 (30.60)

9.07 (21.05)

16.94 (27.04)

16.72 (27.74)

50.71

58.72

47.37

50.00

Involved in the street economy in the past 3 months

45.51

14.75

36.59

36.72

Setting helps avoid street economy

58.02

76.79

68.78

66.42

27.30 (8.91)

27.59 (7.84)

29.20 (7.54)

28.35 (8.10)

School, job training, and work

Substance use Days used alcohol/drugs in the past 3 months (range 0-90) Setting helps manage or avoid drug use

Street economy

Perceived resilience Perceived resilience (range 0–40)

Note: Shaded rows indicate differences among groups at a statistically significant level.

than alcohol or marijuana (data not shown on Table 5). Half of RHY reported the setting helped them to manage or avoid substance use. A little more than one-third of youth (37%) were involved in the street economy in the past three months, although this was less likely among youth in TLP settings compared to the other types of settings. About two-thirds of youth (66%) said the setting helped to avoid the street economy, particularly for youth in TLP or dual/ multi-program settings. Almost all (86%) reported a moderate to strong desire to avoid the street economy (data not shown in Table 5). Results of mixed-effects regression models are presented in Table 6. After controlling for youth characteristics (sex, transgender gender identity, age, sexual orientation, race/ethnicity, time out of home, days in TLP, days in DIC), there were differences in the proportion of youth engaged in school, training, or work by type of setting (χ2 [2] = 6.79, p = 0.03). Engagement was more likely in TLP and dual/multiprogram settings than in DIC-only settings. Youth in downstate settings were more likely than those in upstate locations to report that the setting helped them to engage in school/job training/work (χ2 [1] = 13.17, p b 0.01). Overall setting quality also was related to youth reports that the setting helped them to engage in school/job training/work, with reports of help more likely in higher quality settings (χ2 [1] = 15.43, p b 0.01). None of the setting-level factors were related to recent substance use days after controlling for youth characteristics. In the mixed model for this outcome with only youth characteristics included, the average youth was expected to engage in substance use on approximately 11 of the past 90 days. Setting-level factors were related to youth reports of being helped by the setting to manage or avoid substance use. Reports of help were less likely in downstate settings (χ2 (1) = 6.81, p = 0.01) and were more likely in settings with higher overall quality (χ2 (1) = 5.34, p = 0.02).

Recent involvement in the street economy was related to setting type (χ2 (2) = 6.15, p = 0.05) and setting location (χ2 (1) = 6.76, p b 0.01). Involvement in the street economy was least likely among youth in TLP settings and most likely among youth in DIC-only settings, and also was more likely in downstate than upstate settings. Youth reports of being helped by the setting to avoid the street economy were more likely in upstate locations (χ2 (1) = 10.24, p b 0.01) and those with higher overall quality (χ2 (1) = 9.09, p b 0.01). Perceived resilience was related to overall setting quality (χ2 (1) = 6.24, p b 0.05), with higher ratings of perceived resilience among youth in higher quality settings. A one-unit increase in the overall quality of the setting was associated with an increase of 2.56 in perceived resilience ratings. 4. Discussion The present study advances the literature on the settings that play a vital role in the psychosocial development—and even survival—of RHY, one of most vulnerable populations of young people in the U.S. We studied a diverse set of randomly selected organizations in a discrete geographical area, comprised of rural, suburban and urban regions. These settings evidenced good quality overall, with no differences in quality among DICs, TLPs, and dual/multi-program settings. In fact, none of the settings were rated below the middle of a multi-perspective rating scale, although quality did vary within this range of moderate to high quality scores. We speculate the overall positive quality of settings is driven in part by federal RHY-specific legislation and its associated fiscal and programmatic support, which emphasizes the Positive Youth Development approach, as described above (Levesque, 2011). Yet, as noted above, a substantial number of settings closed or ended their RHY programs during the course of the study. These program closings underscore challenges RHY settings face to maintaining program

M.V. Gwadz et al. / Children and Youth Services Review 73 (2017) 398–410

407

Fig. 3. Overall setting quality.

stability and longevity. In fact, there is agreement among RHY providers and stakeholders that funding levels are generally insufficient to meet the needs of the population; for example, tens of thousands of RHY

are denied shelter each year due to insufficient appropriate housing resources (The National Network for Youth, 2013). These fiscal problems have been even more acute in the period of economic downturn

Table 6 Associations between characteristics of settings and RHY outcomes in mixed-effects regression models.

Associations between characteristics of settings and RHY outcomes in mixed-effects regression models. Engagement in work, job training, or education Coefficient TLP vs. DIC only Dual/multivs. DIC only Dual/multi vs. TLP Downstate vs. upstate Overall setting quality

1.21 0.77 -0.43 0.11 0.30

SE

p

0.60 0.30 0.56 0.39 0.42

0.05 0.01 0.44 0.77 0.48

Odds ratio 3.34 2.17 0.65 1.12 1.35

95% CI 1.08, 12.07 1.14, 4.24 0.20, 1.83 0.50, 2.52 0.54, 3.17

Substance use days in the past 90 days Coefficient TLP vs. DIC only Dual/multi vs. DIC only Dual/multi vs. TLP Downstate vs. upstate Overall setting quality

TLP vs. DIC only Dual/multi vs. DIC only Dual/multi vs. TLP Downstate vs. upstate Overall setting quality

-0.48 0.02 0.50 0.12 0.35

p

0.41 0.26 0.35 0.25 0.27

0.25 0.94 0.15 0.63 0.21

95% CI – – – – –

-1.23, 0.31 -0.42, 0.51 -0.15, 1.15 -0.31, 0.60 -0.16, 0.88

Involvement in the street economy in the past 3 months Odds Coefficient SE p 95% CI ratio -1.24 0.54 0.02 0.29 0.09, 0.81 -0.17 0.35 0.62 0.84 0.40, 1.68 1.06 0.49 0.03 2.89 1.14, 8.06 0.84 0.25